GitLab Support Operations, 27 Oct 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Support Ops - Zendesk - ZD Partner-SFDC Sync

Description

Jason Colyer, Support Operations Manager, takes us through how the ZD-SFDC sync works

A

Hey y'all uh jason support operations manager here at git lab and today we're going to go over the zendesk salesforce sync for partners. Specifically um it's one of the three sync mechanisms we got in place. We've got the one for the main instance, one for the federal instance and uh this one we call partners, even though it also sinks to the main instance. The reason for that is partners are slightly different in how they think so.

A

uh It ended up being easier just to kind of split it off and do its own thing there's a future where we might incorporate it back into the main sync script, but as both get more and more complex over time, it seems more like they'll just be three different entities.

A

So yeah. Let me uh share my screen and we will get started cool uh real, quick mike. Are you able to read this or do I need to zoom in for.

B

That looks great. I can see it just fine.

A

Awesome cool, so um here we've got the repo itself. We've got the basic readme file, it just kind of tells you what it is: uh the required gems, which in this case json oj and restforce I'll, go over those a bit with the gem file, a file tree that shows all the different files and then the soql query that is used and again like I'll go over that more when we get to it in the salesforce file, but the big way that this works is through this gitlab ci file.

A

One thing you'll definitely learn from this video is I'm not original on naming, so everything has pretty bland names. The stages are gather data, compare and sync, and we've got the cache path of vendor ruby. That's to help it speed along a little bit. If you have to retry jobs uh the before script, I haven't output, the ruby version useful for debugging purposes, especially if the ruby image being used from docker updates.

A

For whatever reason this is super useful, just to kind of let you know yeah the instance says it's this version, and then I can check the gem file and other things and say right, but I built it with this. So there's a problem.

A

um We installed the bundler gym and then I'm going to use that gem to run the bundle command, which will then install everything from the gem file, which prevents me from having to type out gem installed blah blah blah blah blah blah blah.

A

The gathered data stage basically has two different parts.

B

A

Precinct zendesk org report and the precinct sf org report. They both run only via schedules. The zendesk org report will gather the zendesk organizations the salesforce one will gather salesforce accounts and they'll both generate artifacts that expire in four hours. The zendesk one generates zendesk.json, which will just have all the partner organizations in it or more.

B

A

More specifically, all organizations in it since there's not a great search term, for that the salesforce will generate two different artifacts we've got salesforce.json, which will contain all our syncable sales force, organizations that are partners and then salesforceproblems.json, which will contain any problems that the gather sage found. That would prevent syncing.

A

The big one would be duplicate names, since it's like, I can't figure out which one's the real org the compare stage is. The precinct. Org comparison report is the job it needs. These two previous gather data stage jobs that way it can run since it's using the artifact files, namely the zest.json salesforce.json, runs only via schedule. It runs the compare organizations and it'll create two different artifacts creates and updates.json, which will contain information about what orgs to create and what orgs to update.

A

And then the final stage is sync that has the sync organization's job that requires the comparison. Job finished and all that's going to do is run the sync organizations, no artifact files required at that point.

A

So I'll go over one of the scripts themselves, just because the rest of them are pretty much the same.

A

It's requiring the zendesk zd sfdc sync partners file and then it's basically calling the module the class and telling it to run you'll find out. The rest of them are pretty much the same. The zendesk one will call zendesk salesforce partners, zendesk gather, run or zendesk sync run, um I'm not original on the scripting there. So it's pretty much just easier to call that than in the scripts. For that ci file, say oh yeah require this and then run this and then run this.

A

It's just a lot easier to call one function and be done with it.

A

Let's go over the gym file. Real quick, like I said I would cool. So uh this is a very basic gem file. It's requiring three gems um json oj and restforce json, being a gem, slash library to handle json objects converting to and from ruby objects.

A

Oj is a nice gem that I like to use. It takes http, headers and requests and stuff like that and makes them into ruby objects, so they're easy to work with, and then restforce is the official salesforce gem to interact with the salesforce api. um The reason I call 4.2.2 specifically is that's the version. I've tested, newer versions, might change api, endpoints or syntax or stuff like that, haven't been tested yet.

A

So, specifically, I know with the ruby version we built this with and the sales restforce 4.2.2 everything works like it should, um as we you know, update. You know as we look at the newer gems, we'll update this accordingly.

A

All right so the zendesk or zd sftc sync partners.rb file, which is the main file. You saw the script, requiring um it's going to declare the module, zendesk salesforce partners, prepare to hear me say that statement a lot um and it's going to require bundler setup and then the bundler is I'm going to call bundler to require default, which is a fancy way of saying I'm requiring the bundler gem and then telling bundler to read the gem file and require everything else. I need this particular script.

A

Only has three gems, so I could have said require json require oj, require rest force, but when you start building up your scripts, you get to where it's like. Oh there's, like 30 gems, I'm using it'd, be really annoying to type out require that many times. So this is a nice little shortcut that just says right require everything you've already installed and then from there I'm requiring relative to call the client file.

A

So the client file would be the next thing to go over and uh once again I've been clearing the modules into salesforce sync partners and then I'm declaring a class called clients. Here, I'm requiring a compare, salesforce and zendesk scripts which we'll go over in a bit.

A

I declare a function called retry options. It just contains a hash that has basically options for telling faraday how to handle certain failures, in this case connection failed and timeout error, and what this is saying is you're allowed five max retries. I want an interval at least of a second. I want some randomness of 0.5, so it'll tweak the interval a bit. So it's not always the same. The back off period between failures is at least two seconds or is uh the backup factor is two meaning.

A

If you get the same error twice, make sure you wait a bit, um but all this will be passed to faraday, to tell us how to handle errors. That way. When we get a timeout or a connection failed, we just don't abort the script instead, it'll say cool. This is what I should do. I should retry.

A

We declare a function salesforce, which just has the salesforce variable if you've never seen this the double pipe equal sign in ruby.

A

I love this because it basically says right set that variable to this, unless it has a value already, it's nice, so that I can call the salesforce variable over and over and over again, and it doesn't make a bunch of restforce objects and instead says right, make it the first time but then just use the same object over and over and over again.

A

I love that feature of ruby. It's one of the things I use, probably the most in scripting, but essentially I'm calling the rest force gym, making a new object. It uses the host the username password clan id client secret security token, and then the api version is 41.0.

A

That's what's been tested with 4.2.2 restforce, um I believe salesforce is actually up to 52.0. So I need to do some testing to see how the new versions differ and all that fun stuff.

A

um We've got the zendesk function, this one's actually calling making a new faraday object, it's using the zendesk url and I'm passing those retry options into it to tell it how to handle retries.

A

I'm setting the the adapter to whatever the default adapter for faraday is um I'm setting the content type in the headers to application json, not as required for get request, but when we do like, creates and updates which we'll use like post input absolutely required because we're going to be sending off json and receiving json. So it's important to make sure that the content tag knows what it's dealing with and then I'm saying the basic auth to be a username token and then the actual api token.

A

um As a fun fact, basic auth is being deprecated. I found that out when I coded this the other day. So eventually this will be updated to use the new method, which will pretty much be the same thing just with whatever new official method to do basic authentication through faraday requests, and then we've got this request function.

A

This takes an http method, url and parameters parameters are by default, set to an empty hash, but this allows me to customize if needed, and all this is doing is doing a public send on that faraday object using the http method and a url which will be appended to that url. We set up here and whatever parameters, I'm passing it's going to take whatever the response is and send the body to oj.

B

A

That it's nice and easy to read and work with any questions on the client file. Before I go to the next section.

A

Cool so I'll go to the next stuff, uh yeah we'll start with salesforce, that's the easier one to kind of start with, so we have salesforce um you'll notice. I'm doing this less thing. Client, that's basically ruby for saying inherent what's already in client, it makes it so I don't have to say zendesk salesforce, sync partners coloncolonclient.salesforce.

A

Instead, I can just say salesforce: it's it's already inherited that it makes it really easy to call to it. We've got this run function which you saw a reference to the script. All this is doing. This uh basically is doing everything. This would actually do so. It's gathering the data verifying it reporting on it and making our artifacts um gather is just printing out that it's gathering date gathering the data.

A

It calls the data function and then it says it's done and the data function is going to call that salesforce object and run a query and we're going to pass this query string we'll go over in a bit. It's going to map those results to account objects we'll go over duplicates in a second in a bit, so the query string. This is what we have here. It's soql, which is salesforce, object, query language. If you know sql you're going to notice, it looks almost the exact same, but essentially what we're doing here is telling it.

A

I want the account id the 18, digit format or 18 character format of it. Sorry, not digit the name, the arr, the sales segmentation or market segment, whatever you want to refer to it as the account owner the number of seats and the license, the type which would be like customer prospects or, in our case partner, um the technical account manager, the support level. The manual support upgrade the region, the game site health score, which, admittedly, I don't fully know how it's calculated.

A

If there's a bunch of factors, gain sight uses, but essentially, there's green, yellow, red, green meaning, it's good, yellow, there's, a warning red, meaning, there's they're.

B

A

Risk my my understandings have something to do with likelihood to renewal, but um I'd say, look into gain sight more to kind of understand how they're doing that uh next renewal date is a calculated feature, calculated field that basically looks at their license. Information says this is when they're supposed to renew next uh partners.

A

uh Partners, partner status, which is weird to say, is essentially a status that will either be they're, active, they're, former they're unauthorized.

A

Basically, it's telling us well, they may be a partner what their actual status as a partner is, and we've got partner track, which will tell us what kind of partner they are we're, selecting all this from the account table or account object as it's referred to in salesforce, and uh we have this. Where statement, we only want partners for the same. We only want it when the status is authorized or.

B

A

And for partner tracks, we only sync open selected technology technology is an alliance partner. It's called technology, not sure why. I just accept that and move on with my life, so um account objects is basically going to generate this hash. That will then be used to help, compare it to zendesk it's kind of putting it in a format that will be easily comparable, which we'll see when we go over the zendesk stuff. But it all starts with this check account which, on main and us federal, does considerably more on this one.

A

It pretty much just determines the type of account and sets the subscriptions ultimate and the support level, because we just default to ultimate for partners select or always ultimate and open have their own stuff that that changes the requester to their customer, at which point the customer's. So you know, plan levels what we actually take into effect and then alliance are always ultimate as well.

A

The type pretty much is just looking at the partner track, and if it's select it puts select partner if open it's open partner or it says, alliance partner. This is just putting it into the format that zendesk expects it in. So it's easier to compare and we don't have to worry about zendesk, saying well. The api name is this in salesforce saying right, but the pretty string humans want to see. Is this um so yeah? That's all that really does, um and then we pretty much just you know. The name is the account name.

A

Salesforce id is the account salesforce. So you know the account id18 for the short id. I just literally take the first 15 characters, which is the short id, is just the first 15 characters of the 18 character, one um for support level. I called out first just in case they have ultimate that's or they have multiple support levels. That's something from main and federal. It doesn't apply here, but I just kept the logic there for sake of simple scripting.

A

Annual support upgrade is always set to false. It's not going to be true on a partner partners. We don't do priority prospects. They have to be authorized already at which point they're not a prospect they're a partner for the gain site health score. I just set it to a string which is, if it's empty. We don't want to send nil, because that's going to do weird things, we want to set it to an empty string.

A

That's what zendesk is going to expect when we pass that on later on yeah, that's uh where that comes from um I've got determined region here, which basically is just saying if it's blank set it to an empty string. Otherwise put it in this specific format. This org region region name to a string dot down case, which is just what zendesk is going to expect a region to say it doesn't expect it to say a merit, expects it to say or region amer.

A

um So the duplicates which we saw up here we've got. This determined duplicates is what it sets the variable to and determine duplicates is going to take all that salesforce data we have, and it's going to group it by the name of the partner in down case that way, it's all lowercase. So it's easier to compare because zendesk is case sensitive or is not case sensitive on naming so bob in all capital.

A

Letters equals bob and lower case letters, and both of them are going to have the same name, and since names are unique in zendesk it will not work. So what we want to do is basically say right, grab all the names that are the exact same, because we can't sync those um this select right here is basically saying right. So of these group. Things only grab the arrays that have a size bigger than one meaning, there's two of them three of them, four of them.

A

Five of them we're doing, flattened just to take this array of arrays and just make it basically one giant array.

A

uh The verified data was is where we're actually going to use this duplicates and all we're doing is taking our salesforce data and subtracting duplicates, because we can't sync duplicates there's no way in the script to say well, this account's right and this duplicate is not so we kind of just go with the. We can't sync them. That's something that whoever is the account owner or tam, or solutions, architect or whoever's, managing whomever, whoever whomever, whichever is managing the account, would you know.

B

A

Of need to fix normally.

B

A

Normally they need to be merged because they're the same account. Sometimes they just need to be renamed either way, we'll fix it from there. It reports on them to let us know, here's how many accounts are located. Here's how many duplicates I found and then it's going to print out all the duplicates with their salesforce id, so you can quickly and easily see which ones are the probable ones.

A

The artifacts it'll generate that salesforce.json and salesforceproblems.json salesforce.json contains the actual data we're going to use for the syncs salesforce problems at this point contains the duplicates that way. We also have an artifact. We can download and use if needed. So any questions on the salesforce script.

A

Cool then we'll go over to zendesk. Oh, I should start with the zendesk file. All the zendesk file does is declares index class require the gather and sync scripts, it's not very original, but it allows me to separate the scripts out. So we don't have a huge file. I can instead split it into smaller, easily digestible files.

A

um You know, we've got the modules index, salesforce sync partners, and then we declare the class zendesk here we're calling class gather. You can see how I'm inheriting from client again and run's, going to generate an artifact and then report on it.

A

So uh the generate artifact is actually going to be calling the data, which is where we're going to gather our orgs and all this is doing, is making faraday requests to gather every organization in zendesk there's not a great way to search just for the partners.

A

um We also don't just want to search for the partners, because if there's sync issues, we want those to be reported later on during the actual sync, but essentially it's going to gather it's going to paginate through every page and as you can imagine, this takes a while. It makes a good number of api calls.

A

It only gets worse the more organizations we have, so that's always fun, but uh it's essentially mapping these out to organization objects.

A

um If there's not a next page in the object, it's gonna break out and then just return orgs, it's a fun fact in ruby. I don't have to say return orgs by saying orgs is that last statement um it knows: hey yeah, so return. The last thing we actually got. um If I didn't have that there- and I just did quits done it- would actually just return true, because it's like right, I I was able to run the command puts done so by saying orgs.

A

It tells it to return orgs and saves me having to type out return. So it's a nice little shortcut uh beyond that. uh The org object. You'll notice. This looks very, very, very similar to what we just saw in salesforce. That's because we want them to look the same, so we can easily compare them.

A

uh The only thing we really have here is we're grabbing the actual zendesk id and the tags this will be used for in partners case not a whole lot in the main and federal it's using it to determine if they have multiple subscriptions and stuff like that, but the essentially support level. um I've got this right here where it's like return starter. If it says basic because basic's not a real plan, it's starter plan. This is completely not needed in partners, because we already know the support level is ultimate.

A

It's always going to be ultimate, even if it's not it'll end up saying I need to update anyway, because it should be ultimate uh determined subscriptions, which I don't like it's there from the main and federal, but you'll notice. It's not actually called up here, because they only have ultimate they're partners. It's a little simpler. I just didn't delete the function.

A

um Clean tags is going to duplicate the tags. It's going to remove all the plan tags from it, and then it's going to sort them out. This will be used later when we compare them to make sure they need to be synced or not um again like we do it for this one. But it's not super needed with partners, because at the end of the day, like there's, not starter and premium, there's not bronze silver gold, community custom, they're they're, only ultimate yeah, so that's gathering all the zendesk or any questions on that. One.

A

Awesome so we'll go back to compare here's where we're actually going to compare them and start getting ready for the sync. We declare the classes. Compare we don't need to inherit from client because we're actually not making any api requests at all. We're just literally running some ruby on this one. We've got the run function, which does compare generate artifacts reports.

A

I set up some variables here for myself. I've got creates, updates, do nothing. This is mostly just so. I can have a running tally so that we can report.

B

On it generate.

A

Artifacts, based on what we need, but uh the compare is essentially going to go through every salesforce account. That's what accounts is, I believe. That's declared down here at the bottom organizations is coming from data zendesk json. Our accounts come from salesforce.json, but essentially we're looping through each account, I'm doing it with index.

A

This is solely so I can have the script telling me what number it's on, because this does take a little bit of time and I have a little quirk where, if I see a script saying it's running and then nothing happens for like two minutes, I start panicking that it's frozen. This gives me output pretty consistently. So I know it's constantly running um and then basically, we're going to in the zendesk works.

A

Try to locate the one where the salesforce ids match, because that's how we sync them is based on salesforce id if it can't find it in zendesk. That means it needs to create it. So it's going to push a create object to the creates array.

A

Otherwise, it's going to determine if an update is needed unless it couldn't find it because we already know it needs to be created, so uh your create object. It looks like this and this is starting to move towards the format. Zendesk wants it in. um It wants a name and then it wants organization fields and then it wants each of those listed and it wants the tags.

A

If you don't do it this way, it'll annoyingly tell you hey, I did it and then do nothing at all, because you passed a bad object and it's like right. I sent that to be created, nothing was created because that was a bad object.

A

um The com determine update it's going to make a comparable object, which basically is removing those objects. We can't compare like there's, not a zendesk id in salesforce, there's, not tags in salesforce, so you'll see it's essentially removing those it's ensuring that health score is a string because we really want to compare it as a string, and then it's literally just saying: okay, so does that account object equal our new, comparable object.

A

If it does cool, nothing has to happen. It's already in sync, everything's, perfect: it's going to push that to do nothing, so we have it stored there. If it's not the same, it's going to actually push an update, object to the updates array, which looks a lot like the create object, except it has an id.

A

That's that's pretty much the only difference it has an id and we also call the update tags there, which is essentially just going to duplicate them, ensure that that ultimate tag is there and then it's going to do a sort and unique, don't actually have to sort them here, but it's great for debugging. I like things in alphabetical order. So that's why there's a sort it's not required.

A

You can push them out of alphabetical order, but it's a lot easier for me to read them when I have to debug or troubleshoot, if they're already in alphabetical order, if you're not familiar with the dupe or dup. However, you want to pronounce that function.

A

It's a fun function in ruby that says so clone that and make a whole new one. So I can edit it and do things to it without impacting the original. What this does is allow me to edit the tags without actually editing that org tags object super useful. That way. If I need to debug, I can come in here and say: yeah.

B

A

Out what the tags were print out, what our new tags were that way, I can make sure they're what they need to be. uh If I just mingled the tags object, then I have the problem of. I have to output it before the main goal, whereas here I could say, yeah push here and then output both after that still have what I need, but yeah beyond that uh we've got the report, which is going to tell us hey, I'm creating. I need to create this many update this many.

A

This many were in good standing and then we'll generate the creates.json.updates.json from our creates and updates array. So any question on compare.

A

Cool, so our final step of this is the sync script, so you'll notice here it's you know, class sync, I'm inheriting from client, so I'm going to be making api calls to zendesk. I've got my run function, which does this created, update errors, create or sync towards a report.

A

I shouldn't actually need these. This is actually duplicate coding that I could probably remove, but all I'm doing is creating these create, update and errors, objects created and updated, or just the number zero. That way I can do a quick valley count based on successful and errors is an array that way I can push a bunch of errors there and at the end, have nice output that says hey.

A

I failed to do it because of these reasons so uh create orgs is just gonna go through that creates, which is you know that creates.json we made and it's gonna do each with index. I like output, so there's output randomly telling me what number it's on and it's going to create. The org, which is just gonna, make a post request to organizations.json using a create object.

A

If there's an error in the response, it'll put that into the errors array in an error string we can use later. This is the create object which is the ultimate. This is exactly what zendesk wants has to be in this format. It wants you to say it's an organization and then go to a hash and say the name and the tags, and then say the organization fields and go to a new hash, and this is where, if you're slightly off it will not do things.

A

It'll actually maybe do annoying things like blank everything out, and the last thing you want is to create the org, but have it all blank because then you can't create it later, because it's going to air out and say that org already exists so end up yeah at the end of that hash, we're putting that into a giant json string, because that's ultimately, what it really wants is that format in a json string.

A

uh The updates um we'll go down here to sync orgs and go back to that object there. But here I'm going to instead of doing it one at a time I'm going to actually do a bulk update and senddesk has a limit. You can only do 100 of those at a time, so I'm taking the updates.json list and I'm slicing it by 100, which is to say, give me a hundred. Give me arrays that contain a hundred of each 100 items before you go to the next array um from this.

A

Essentially, I'm making this giant sync up, I'm making this sync variable, which is going to have organizations and then it's going to map all the orgs to the update object, which looks a lot like the create object, except it has an id and tags because we don't need tags on it newly created, but we do want to make sure they're updated on existing ones.

A

It's going to send the request and then it's just going to output the job status. That's because when you do a bulk update, it's not going to say hey, I'm successful because it's got 100 or so orgs to update. Instead, it's going to say, I'm doing it via this job id.

A

So we have it print that out in case we need to debug. We can go look at that job status and that will actually contain a bunch of information. Saying I succeeded, I failed. I succeeded, you know this is why I failed, but essentially we're just going to count. However, many orgs are in a slice normally a hundred, but you know if we have like 579, that last iteration you can only have 79 so we'll add that to the updated and uh yeah after that, we're just going to report them. So we've got.

A

I created this many. I updated this many.

A

If there are any errors in the array, it's going to actually print out the error, count and then print out an error report which is just going to print out a bunch of error strings and all that is, is this format right here saying hey this salesforce id could not be synced. This is why 99 of the time, the reason why is duplicate name could not sync, there's already an org with that name, something's wrong um and usually with uh at least with updates or creates.

A

When we see that our ops has to go and find out why it exists, but it doesn't have a salesforce id and then fix it so yeah any questions on the sync scripts.

A

Cool, so um that is the basic gist of the partner sync. All of this runs via schedules, uh ci cd schedules, which I believe, run every 30 minutes. Currently, um we've got this in the ops. It's because this is going to output customer data we needed to make sure it was as access controlled as possible.

A

That's what we've got on this ops instance where the admins are very tightly watched and only specific people are allowed access and, like as an example, is partner. Sync, I believe only support ops and the admins of the opposites can see this. I have plans in the future to figure out how to sync or mirror the scripts over to.com so that they're visible, but not running, but yeah. That's uh the basic gist of that.

A

Any questions. Comments concerns thoughts, any of that fun stuff.

B

It all looks good to me thanks for putting that together, jason.

A

Oh yeah, no problem at all. um Like I said it's, you saw it's different enough that I don't want it in the main sync script. The main sync script is complex enough. um That's why we kind of have it here because it's allowed to skip so much of the checks, because it's always ultimate. It doesn't matter. If salesforce's salesforce says well, it's bronze doesn't matter. It's a partner partners are always ultimate. So that's why it's separate. I like that.

A

It's separate this is my favorite one to work on, because it requires very little thought to do.

A

Right on uh any other questions, comments, thoughts.

B

When you're looking at the duplicates, um those are duplicates within salesforce itself,.

A

Correct um good examples: if you uh a good one I can point out, is there are about 100 accounts in salesforce called test, so.

B

A

What it ends up doing is in its report for at least maine, where that's a where that's an issue. It has about 100 lines. That say, I found 200 duplicates of the word test and it just prints them all out. um I've confirmed with sales, ops and sales in the past, so things with that nature of a name will not be synced because they're supposed to be test accounts.

A

um There could be a point in the future where there is a company called test, and there is. We have to fix that that at which point I'll basically be like there's 99 of these, you got to get rid of.

B

And what do we do with the more legitimate duplicates.

A

I tend to check them once a week at minimum, but usually a little more and I tend to generate a report and send it over to sales saying hey. None of these are going to sync because of duplicates um and let sales act to you know action on those as they have time and effort to do so if it comes up in an actual ticket like a ticket and somebody's like hey this didn't sync, you don't have sla.

A

What's going on, that's the point when I tend to go to salesforce and make a chatter and say hey, we have to fix this because there's a legitimate customer filing a ticket and it has now impacted their ability to receive support.

A

uh There's- and I know, there's tooling they're trying to build in salesforce to prevent the duplicate naming issue, because it's not just zendesk that has a problem with that. There's several other tools. We use that do not like things having the same name.

B

I assume you have something similar for synchronizing the contacts people.

A

Unfortunately, not in global, um unfortunately, global has over 300 000 contacts. So to get a list of contacts in zendesk last conservative effort fastest, I could code fastest, I could get it to run. It took six and a half hours.

A

Unfortunately, we cannot run a ci cd job for six and a half hours, uh whether it be in ops or gitlab.com, which you cannot do that um I am actively working with zendesk. I've got a couple feature: requests of like I'd, love to be able to do this and not need six hours.

A

I'd love to not have to iterate a hundred contacts at a time um federal has a contact sync, it does the basically the same thing you just saw, but with contacts with the caveat and it's something I've addressed with like james and federal is at a certain point. We will hit that threshold of it's too many contacts. We have to stop, which will then require some process changes um with zendesk main. We were already past that threshold on like day one when the script with a very first v1 of the script was made.

A

It is something I've tried many times to find out a good way to do. I've got it down to about six hours, which is great, considering the original I had was like 12 and a half hours, but six hours is still way too long.

A

uh That's why we have like the zapier stuff in place to do that, like auto association, stuff, that's why we have, uh and if the official zendesk, salesforce, sync and zendesk integrates with salesforce, for when contacts are updated to add them which works mostly.

A

um I say that because it does have errors every now and then, especially when it's like. I can't put this guy in two different organizations, and it's like correct you cannot. We do not have it set up to do that, but uh it does help at least eliminate some of that. Unfortunately, we do encounter cases where they're, not in salesforce or it didn't sink them properly or couldn't associate that's why we have a needs.

B

Okay, thank you that helps to understand it quite a bit.

A

Yeah no problem, anybody else, questions comments or might can you can throw out another one if you want.

A

Cool well I'll call them I'll call this all that I appreciate you joining. I hope this was educational. I hope you learned something if you ever have questions or want to talk about this feel free. To put some on my schedule, come to the ops channel dm me, I'm always willing to talk about the six scripts.