GitLab Support Operations, 3 Sep 2021

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Support Ops - GitLab.com Projects - Customer Ticket Generator

Description

Jason Colyer, Support Operations Manager, takes us through how the Customer Ticket Generator works.

A

Greetings all my name is jason coleman support operations manager here at jet lab and today we're going to cover the support, the support, ops, customer ticket generator and how to work it, how it works so on and so forth. So let me share my screen and we'll get started.

A

Okay, so what we have here is the customer ticket generator project, and essentially all this is holding is the html files from the form itself.

A

This is just a basic html form asking for the required information towards the bottom. We have some basic javascript checking for you know bad entries and missing values. Beyond that, um it will then submit. You know an ajax request to adapt your web hook, which will get us in the place. We need to be the zapier web hook, which we can see here um at the top we're importing requests in json.

A

This isn't required, but it's something I've gotten into the habit of doing it doesn't hurt to require import them, but it does hurt if they're missing. So I was there on the safe side here.

A

The first function we see here is the find user by username. This is essentially taking the username. They submitted the one they said. This is who I am and then it's double checking it's sending a request through the getlab.com api, to check if the username that they, basically they can be located in gitlab.com.

A

uh This is so that we can do stuff like make sure we put their username and all that fun stuff into the form and make sure we ping the right person.

A

uh The next thing we have is the ticket label. This is basically going to take the issue link. They give split it up to grab the uh the project it exists in and then it's going to add the issue id to it. We also have the ticket body function. This is basically taking the description they gave, but replacing name space list here with a namespace list.

A

A namespace list item is basically just an asterisk and the item itself. uh The namespace list is going to take the namespaces from the user based on the csv they've. Given us, it's going to split it by space and it's basically just going to give us a bunch of user item a bunch of namespace items.

A

um These two functions here work to uh basically not only parse the csv string but double check it, namely it's going to split it by lines. It's going to look and say hey. There should be three columns: three comma delimiter, comma, separated columns.

A

If there's not something's wrong with the csv file, we're going to return false, we can say it's not um if it's you know uh beyond that, we're going to also make sure for actually parsing it we're basically just uh doing split lines on it to say, yeah split, each of this giant string by new y.

A

um Here we have the failed issue description. All this is doing is adding a bunch of information to generate the issue to say: hey, it was a failed request. Here's why we can't proceed. Success is doing the same thing.

A

This giant function here and I'm not going to go too much into detail over it. We'll go over the detail of it by looking at the actual script in a bit, it's just generating a ruby script for you to run and use in the actual process. This is to make your lives easier so that it's quicker to kind of process. These you don't have to use the ruby script generated. You can absolutely manually, go make the tickets or do something of your own design.

A

uh The key here is: we want to make the tickets we want to make it right. We want to make sure we know what issue link our ticket links. Were you know what the link to the tickets are, so that we can add them back to the issue so we'll skip through this labels that it's always going to add on the issue here: um support ops category other sports party normal. This is for zendesk global and it's a support, ops to do item uh for a failed request issue.

A

The title is going to be customer ticket generator failed request. The description comes from that function above the labels are the ones that we got here and it's going to create the issue. Success is basically the same kind of deal. Customer ticket generator successful request the description that we generated above the labels. It's going to assign the username of the user who submitted this ticket. That's so that they get updates on this, and then it will create the issue.

A

Creating the issue is just going to make a post request to the projects api that were for the issues api for the project that we're keeping these in um the content. Types: application, json, the private token, is a get lab token to generate these um list. Submitted values is for a failed request so that we can see what they submitted in the issue itself, so that we can kind of point to it. If they ask questions but beyond the functions we get to the actual script itself, we're setting an empty array for errors.

A

The subject description, submitter are all coming from the input that they gave. What basically what they put on the form. um The author is the submitter. We do that just to duplicate the variable, so we can modify it as needed. uh Csv is from the csv content that they gave us uh we're calling strip on it to get rid of, leading and uh trailing white space or yeah. Sorry, white space is beginning in string um issue. Link priority and token are all coming from input data as well, whereas issue link priority are coming from.

A

The form. Token comes from a setting that we have in zapier the user, we're calling that find user by username the entered values we're getting that listed, submitted values.

A

So, basically, here, if it can't find the user, it's going to add to the errors that it's invalid getlab.com username and the author is going to be set to false if the csv doesn't pass the checks, it's going to say it's an incorrect format from here we're going to try to get the issue labels and the reason for this is we have this attribute index error.

A

These will occur if what they gave us is not in a valid format, and we want that to fail, but we don't want the whole script to just exit because it fails so we're actually doing a try accept if it hits this it'll. Add it to the error error list.

A

So if there are, if uh the length of the errors is not zero, basically we're going to submit a failed request, but if it is zero, that means nothing failed. We can submit the request normally now. uh This is what the form actually looks like that's generated from the project itself. It's basically just asking: what's your username? What's the subject, what issue is most relating to? What's the priority, the ticket should have what's the message.

A

The contents of the csv file- and it's got some like notes here- to kind of help them, along with the generation.

A

Now, uh let's go into an actual issue and we'll get to one of these test issues that we generated.

A

So I can kind of show you what the script looks like. So here's what the ruby script that it generates for you looks like um it's utilizing the zendesk api. It's setting up a client variable that is basically a zendesk api client object.

A

Taking in you know the url, the username, the token and saying it wants to do a retry. The subject is going to be what they entered. You'll see it like this in a string format, that's something you want to check to make sure the string is valid because if they put like single quotes in it, it might not render correctly or if they put weird symbols in that can't be parsed correctly. That might also cause issues uh the description string.

A

um Basically, all this is doing is giving us the description string itself, the namespaces it's going to list out the namespaces and it's going to add them. You know it's going to basically join them with a new line.

A

uh The ticket object. This is where we get the actual ticket object itself. This is what's going to be sent to gitlab.com to zendesk to generate the ticket. Now um you'll see how some things in here look weird. This is stuff we want to know like as an example, the user's array is empty. That's that's strange, even though this says it's successful, we couldn't actually process this. It wouldn't make any tickets. So this is something you want to bring up. We'd want to kind of go check.

A

Staff here see what they submitted for the csv, come back to them and say: hey here's why this didn't work you can see here. How, like the join, looks weird, that's something we also would want to address. You can see how I have no idea what I'm doing. It's, probably not what they want to be sending to customers. So it's probably worth talking to them and saying hey you. You said this is what the description of the ticket is, but we probably shouldn't send that to the customer.

A

These are uh good things to check, for we want to make sure all this looks right, but in theory, if everything looks good in the script, you feel comfortable running it. You can just copy this on your computer and run it and from there it will generate tickets as needed.

A

That's all there really is to how it how the customer ticket generator works. As far as working the issues we kind of talked about this of like go over the request itself, make sure it looks good, assign it to yourself.

A

um You know, add any tags that you need to add, such as, like the support ops, the academy you know, support ops doing instead of to do because you're now working it um from there, you know review the script once you've run the script grab the output from the script or, if you manually, make the tickets get the ticket links paste them back in the issue in the issue, so that it's tracked add the time that it took you to track it close out.

A

The issue add that label support ops completed, that's really how you work them, but uh if, at any point you don't feel comfortable process absolutely reach out to your fellow support, ops, team members reach out to myself reach out to another support, ops. You know manager we're always more than happy to look over with you and go through the process.

A

That's all there really is to the customer ticket generator. I hope you found this educational and I look forward to seeing the next video.