/
/
Decide whether to keep rolling out a build, on evidence
For engineering teams who want a Bugsnag release stability verdict and the errors driving it, from their AI assistant.
On this page
Connector
Connect BugSnag, pick your AI tool and paste the prompt.
The problem
Rollout calls made on misleading stability numbers
Deciding whether to keep rolling out a build means working out release stability from session counts and finding which errors the build introduced. Builds with only a handful of sessions make the percentages misleading.
Without PopdockAI
- Work out stability per build from session counts
- Compare each build with the project's targets
- Trust percentages from tiny samples
- Guess which errors the new build introduced
With PopdockAI
- Stability is worked out per build, tiny samples set aside
- Each build is compared with target and critical stability
- New errors are separated from older ones still occurring
- Keep rolling out, hold or roll back, read-only
The prompt
Use this prompt
Copy it into any AI tool connected to PopdockAI. Swap the values in braces for your own.
**Objective:** Report on the health of the recent releases of `{project-name}` in the `{release-stage}` stage (default production) and say whether the latest build is safe to keep rolling out. Report only.
**Step 1: Pull the releases**
Call `bugsnag-get-release` with `project-name` set to `{project-name}`, `release-stage` set to `{release-stage}` when one is given, and `max-records` set to 20. Keep, for each build: `app-version`, `release-date`, `errors-seen-count`, `errors-introduced-count`, `total-sessions-count` and `unhandled-sessions-count`.
**Step 2: Work out the stability of each build**
For each release, express stability as the share of sessions that were not unhandled, from `unhandled-sessions-count` against `total-sessions-count`. Ignore builds with too few sessions to mean anything and say which ones you ignored.
**Step 3: Compare against the target**
Call `bugsnag-get-project` with `name` set to `{project-name}` and read `target-stability` and `critical-stability`. Mark each build as above target, below target or below critical.
**Step 4: Attribute the regressions**
For the latest build, call `bugsnag-get-error` with `project-name` set to `{project-name}` and `last-seen.from` set to that build's release date, ordered by user impact, with `max-records` set to 25. Name the errors most likely introduced by it, using `first-seen-date` against the release date.
**Step 5: Report**
Give a table of builds with release date, sessions, stability, errors seen and errors introduced, then a verdict on the latest build: keep rolling out, hold, or roll back, with the reason and the errors that drive it.
**Rules & Guidelines:**
* Read-only. Do not call any tool that changes an error or a project.
* A build with very few sessions is not evidence. Say so rather than reporting a misleading percentage.
* Distinguish an error first seen in this build from one merely still occurring in it.
**Validation:**
* If there are no releases for the stage, say so and list the stages that do have releases.
* If `target-stability` is not set on the project, say that and report the raw numbers without a verdict against a target.Who it's for
One answer, three teams that use it
Release managers
Decide whether to continue a rollout
Get a keep, hold or roll back call on the latest build with the reason behind it.
Should we keep rolling out the latest production build of Fabrikam Checkout in Bugsnag?
Mobile developers
See what your build introduced
Find errors first seen in this build, separate from older ones still firing.
Which errors did the newest release of the Contoso iOS app introduce in Bugsnag?
Engineering managers
Track stability against target
Compare recent builds with the project's stability targets in one table.
How do the last five Northwind Web releases compare with our Bugsnag stability target?
How it works
What PopdockAI does when you ask
You name the project and your AI assistant measures recent Bugsnag releases against your targets. It only reads, and gives no verdict when no target is set.
01
Ask
You name the project, and optionally a release stage.
02
Measure
It works out stability per build, setting aside tiny samples.
03
Compare
It checks each build against target and critical stability.
04
Report
You get a verdict on the latest build and the errors behind it.
What PopdockAI provides
The tools behind the answer
Ready made tools for BugSnag and one secure connection to your AI tool. No code, no exports.
bugsnag-get-release
bugsnag-get-project
bugsnag-get-error
This report is built from Get data tools only. PopdockAI applies the row, field and parameter rules your admin sets, and every query is written to the audit trail.
The outcome
Rollout decisions made on evidence
Related use cases
More ways to put this data to work
Assign Bugsnag error to owner
Ask your AI to assign a Bugsnag error and it finds the person among the project's collaborators, warns about pending invitations and current assignees, and assigns after you confirm.
Learn More
Bugsnag release impact on Zendesk support
Ask your AI about a release's support impact and it compares Zendesk ticket volume before and after, and ties tickets to errors first seen in that Bugsnag release.
Learn More
Give one owner the error and its issue
Ask your AI to assign an error and its issue to one person and it finds them in Bugsnag and Jira by email, shows both current assignees and assigns after you approve.
Learn More
Match Bugsnag release regressions to GitHub commits
Ask your AI to match a release's regressions to commits and it lists errors first seen in that Bugsnag release, the GitHub commits shipped since the last one, and leads per error.
Learn More
Raise an issue for a Bugsnag error
Ask your AI to raise an issue for a Bugsnag error and it gathers the impact, drafts a clear title for you to confirm, checks none exists and raises exactly one.
Learn More
One prompt template with infinite possibilities.
Your AI assistant adapts its approach, tone, and capabilities to match each unique context while you maintain a single, manageable workflow.
Join forward-thinking companies that are transforming their operations with intelligent, secure AI automation.