Automatically refuse L&R STARs that are submitted too far in advance of SLA expiration

Problem Statement

What is the problem?

A high percentage of submitted STARs are for L&R, and over 50% of the L&R STARs are invalid.

Why is this a problem?

We are spending a lot of time dealing with invalid L&R STARs when we should be putting our attention into valid STARs and other high priority needs. The invalid L&R STARs create unnecessary and time-consuming work for on-call managers, and have a negative impact on our efficiency and results. They also create a bad experience for our partners in Sales and CS.

Proposal

Eliminate over 50% of the invalid L&R STARs by automatically refusing those that are submitted more than an hour prior to when the SLA clock on the associated ticket would expire. Details:

  1. The automation should occur at the earliest possible point so that no STAR post is made in Slack and if possible no STAR issue is created in GitLab.
  2. The automation should notify the submitter of both what and why:
    1. What: We have automatically refused this STAR.
    2. Why: As noted in the handbook, a STAR should be submitted only when the SLA is about to be breached or has already breached. It's important to allow the Support Team an opportunity to receive and work on each ticket according to normal work processes.
  3. The automation should also provide options to the submitter:

    If this is an emergency situation, please use the emergency process instead of the STAR process.

DRI

@mdunninger will act as the DRI for this issue.

Required Resources

  • I will need to work with @gitlab-com/support/support-ops to determine the feasibility of the requested automation, and to develop and deploy it.
  • I will need @gitlab-com/support/licensing-subscription to review, improve and accept this proposal

Potential Roadblocks/Things to consider

  • The automation may be difficult or impossible to create within our current toolset
  • This may receive significant pushback from Sales or CS

Desired Outcome

Following the implementation:

  1. On-call managers should observe a significant decrease in the number of L&R STARs received
  2. On-call managers should observe a significant decrease in the percentage of L&R STARs that are invalid

How do we measure success?

Starting immediately upon solution implementation, we should receive 0 L&R STARs more than an hour prior to SLA expiration of the associated ticket

Where would future feedback go?

In a new issue

Related Issues/MRs/Epics/Tickets

  • #4775