SLA Alerting in Minutes

I'd like to add to this, no different from above really

We have several SLAs for initial Ticket Response which are measured in minutes, from 15 minutes upwards.  The current Hour whole based Conditions don't offer what we need.

Can we have "Minutes until next SLA Breach" or allow decimal for hours, 0.1 = 6 minutes would be ok too.

Is this issue caused by data; Integer vs decimal OR by the process that Polls the Automations?

thanks, Mark

Just a quick note to probably set expectations from what I came to know the product - Zendesk has a built in refresher, a task that runs in the background and indexes everything, checks for triggers and fires off events, which runs once an hour. We don't know when exactly, as dev ops may fine tune timing to make sure there are no collisions on pods. This explains hourly triggers (roughly an hour, we don't how much time a task takes), in Zendesk Support as well as Explore.

In order to have a granular per-minute trigger, there needs to be a task that runs every minute and scans the triggers DB table vs the data. It cannot be the refresher (or whatever its true name is), as that is a huge one. So the request is to build a new engine that runs every minute and performs a subset of that task.

We, as customers, do not know how taxing this is on pods and what is the toll on performance, for a per-minute task.

Please don't get me wrong, this is a great product, and it's top-tier. I do think that engineering needs to be creative in finding a solution for this. Maybe spin up shadow pods that their whole purpose is perform tasks (split it from the UI). The bottom line is that other products have this feature, and this is what customers expect to get. Especially when dealing with SLA, which can have financial impact, as the gentleman posting above noted.

Thanks.

Hello,

This is another request to add the ability for SLA breach to automatically do something straight away.

Like many others, the company I work for relies on having an alert sent to PagerDuty if a SLA is breached. Having tested the system and having tried multiple ways to try to get this functionality with what Zendesk offers, I have finally reached the point where I now need to look at moving to another product. 

Now I understand that it's extremely taxing on a system to poll triggers every minute, for 1000s of instances of Zendesk, but surely it's easier enough that on SLA breach, the system can do "something". You're already tracking the time of an SLA and you already know when it's about to breach. 

Honestly, it doesn't seem to me like Zendesk has really thought out this process at all. What's the point of having an SLA in the first place if you then wait an hour before anything happens? 

Adding a comment to bump this. I'm even trying to get creative with automations to get a trigger to go off when a ticket breaches and I'm hitting the "automation can only go off once on a ticket" error.

Please explain how "Ticket: Hours Since Created / (calendar) Less Than / 1 hour" can hit more than once on a ticket....? A ticket can only increase the hours since it's creation. It can never be created for less than an hour more than once if Zendesk automations only check once an hour already.

If we cannot set any triggers up based on SLA or Queue, there HAS to be a way to identify tickets where the SLA has breached and send out a notification to the team. Maybe create a "ticket_breached" tag as a Zendesk Feature? or create a way to send out notifications based on when a ticket enters a specific queue?

Hi Everyone!

Thanks for all your comments and input on this thread. I'm a new Product Manager here at Zendesk, and one of our priorities for the year is to improve our SLA functionality. Thanks for your patience, and continued feedback! We truly appreciate it.

This feels a little like shouting into the void at his point, but I will add that our support team also needs more granular SLA timing.  We have a 2-hour SLA, so being notified halfway through is not especially helpful for us, and paying extra for an addon that should really be natively supported is a tough pill to swallow for those budgeting.  Zendesk, we would love to hear if you have a plan for this.

Hi Dave,

I did see that, but the four-month-old comment about SLA functionality being a priority this year doesn't inspire much confidence in a five-year-old thread as we move into the fourth quarter of the year.  Are there any updates?

Even once every 15 minutes or 10 minutes would be a massive improvement for a lot of us.

I would like to echo Aaron's comment/request for an update. We have a 3 minute and a 10 minute SLA alerting need and the SLA Timers app is causing other issues on our tickets so it doesn't fully solve our needs. 

@Sara Ledger may I know what issues you're having with SLA Timers app?  I was just looking at possibly implementing it in our instance, but I'm curious to know the side effects.

@Cesar Jimenez We're running into race conditions with tags being removed if the action adding a tag is occurring at the same time as the Timer action. I've reached out to Zendesk who said it would need to be handled by the Developer and I reached out to Sweethawk who said it should be handled by Zendesk and that I was only the second customer to report this issue.

With the race conditions, the tag will just be removed but you won't even see a removal event in the ticket event history and you have to assume its based on the timestamps of two similar actions.

We have several automations/triggers that run off of other tags and we are in an emergency industry so the fact that tags are being removed is impacting ticket workflow, agent response, and reporting. Makes it very difficult for other other processes to run smoothly with the tags removed and our Agents questioning what they did wrong on a ticket when they didn't do anything wrong, it was the Timers app throwing off the flow. 

We've now kicked off our project to enhance SLA capabilities and have a team dedicated to this. It's an ambitious roadmap, and we'll be shipping in phases. As things progress I'll share updates, but near real-time notifications is a significant part of this initiative. For now apps like Sweethawk, Cloudset or Myndbend are still the way to go, but sometime in 2022 we'll have this natively. 

Thanks for the update, Scott!  Looking forward to hearing more as things progress!

@Scott Allison Thank you so much for the update!

@Cesar Jimenez and @Sara Ledger, Pete here from SweetHawk - Just chipping in regarding our SLA Timers app. We carried out thorough testing on this after this was brought to our attention and found that the issue is not related to this app or apps in general. It is an issue with the way Zendesk tickets allow updates. If multiple updates to a ticket are attempted at precisely the same time (ie. in the same millisecond) then one of those updates will not materialise on the ticket, however, in the events log, Zendesk will show both updates. You can test and confirm this by setting two triggers to update the same ticket with a webhook at the same time. 

The good news is that it is incredibly rare for this issue actually to happen in the wild because Zendesk seems to have measures in place that try to prevent this and cause errors to show if updates close together are detected. It’s just, very occasionally, those measures don’t work. Nonetheless, if this is of particular concern for any workflow or any app that makes a change to a ticket, you can setup a backup workflow. For example, in the SLA timers app, set an extra timer to go off a minute after the first. That being said, the ideal solution here would be for Zendesk to address the underlying update issue - no matter how rare its occurrence is. I hope this help shed some light on things.

@Peter Godden - It would be nice if Zendesk and Sweethawk got together to resolve this issue since both companies are passing the issue off to the other. I have open tickets with both that specifically direct me to reach out to the other company.

We have so many tags and layers of automations/triggers that your workaround is a large effort to put in place and honestly not very feasible for a normal workflow, especially in emergency situations that my company deals with on a regular basis where updates are happening very quickly and we don't have time to check to make sure if an app is wiping out other needed data.

Additionally, we are not seeing this happen on a rare basis. We are seeing this happen regularly.

Dear Sara and others, SweetHawk is still working on this issue and we are collaborating with Zendesk. It's not consistently reproducible so it's been challenging. We'll certainly keep everyone posted.

Quick update: we've had word from Zendesk this is going to be addressed by one of their core teams, however they could not advise an ETA at this time.

I would like to add that I was very surprised to find out that this feature is not available. We would like to be able to create a trigger once an SLA has been breached so we can send some type of notification to the customer and alert a lead or manager to respond. Thank you!

 @Scott Allison @dave12 @adrian36 @shawna13 an ETA on when the granularity of SLA timer notifications will be implemented is of the utmost importance. 

Considering this was originally posted in 2015 and has been a sentiment of many users, how does Zendesk intend to meet this pressing need?

Thanks everyone for your patience and support! Last year I provided an update that SLAs were a top priority, and we had plans for a number of enhancements. I can confirm that work is in progress, and the very first customer facing change will start to roll out next week: an easier way to see upcoming SLA breaches.

The next big change will be the ability to put OLAs (internal SLAs between teams) on tickets, and that will come sometime in Q3 this year.

After that, we'll be working on this request, to alert on SLAs in minutes. Instead of being tied to Automations once an hour, you'll be able to set up alerts for SLA notifications in near real-time. 

I'll share more details later this year as that work begins, but in the meantime - as others have mentioned below - if this is functionality you need now, check out some of our partners who offer this today.

@daniella13 Thank you for bringing this up! I replied yesterday but I just wanted to highlight it in case you missed that. It's pinned at the top of page 1.

We really need this functionality. We have our first response SLA set to 30 minutes, and having the ability to be notified of near breaches would really help. 

Any update on this issue?