[Developer Documentation] Illustrate the Agent upgrade process #2858
Conversation
ycombinator
added
documentation
ycombinator
requested review from
cmacknz,
juliaElastic and
leehinman
and removed request for
pchila
💚 Build Succeeded
Expand to view the summary
Build stats
❕ Flaky test reportNo test was executed to be analysed. 🤖 GitHub commentsExpand to view the GitHub comments
To re-run your PR in the CI, just comment with:
|
docs/upgrades.md
Outdated
| UW->>A: Start | ||
| A->>FS: Ack failed upgrade | ||
| FS->>ES: Update Agent doc in `.fleet-agents`<br />set `upgrade_status` = null<br />`upgraded_at = <now> | ||
| Note right of ES: Need to check |
There was a problem hiding this comment.
TODO: check and remove note.
docs/upgrades.md
Outdated
| FS->>ES: Update Agent doc in `.fleet-agents`<br />set `upgrade_status` = null<br />`upgraded_at = <now> | ||
| Note right of ES: Need to check | ||
| UI->>UI: Show Agent status as "???" | ||
| Note right of UI: Need to check |
There was a problem hiding this comment.
TODO: check and remove note.
There was a problem hiding this comment.
Agent status goes back to Healthy on successful rollback (by clearing upgrade_started_at field)
docs/upgrades.md
Outdated
| A->>FS: Ack failed upgrade | ||
| FS->>ES: Update Agent doc in `.fleet-agents`<br />set `upgrade_status` = "failed" | ||
| UI->>UI: Show Agent status as "???" | ||
| Note right of UI: Need to check |
There was a problem hiding this comment.
TODO: check and remove note.
There was a problem hiding this comment.
I don't think we show anything, that is this bug essentially #2508
There was a problem hiding this comment.
Yeah I think the agent status stays in updating. I think in some cases the error is written to .fleet-action-results and the UI shows the action failure in Agent activity flyout.
|
Moving back to draft status until the three TODOs are resolved. |
docs/upgrades.md
Outdated
| A->>A: Rexec to start new Agent artifact | ||
| A->>FS: Ack successful upgrade | ||
| UW->>UM: Remove | ||
| FS->>ES: Update Agent doc in `.fleet-agents`<br />set `upgrade_status` = null<br />`upgraded_at = <now> |
There was a problem hiding this comment.
also set upgrade_started_at = null
There was a problem hiding this comment.
There is also a step here where FS writes the successful ack to .fleet-actions-results. We have seen an issue where the write fails (e.g. ES connection error) and the action stays in progress in the UI forever.
|
Pinging @elastic/fleet (Team:Fleet) |
|
Pinging @elastic/elastic-agent (Team:Elastic-Agent) |
ycombinator
force-pushed
the
doc-agent-upgrade-comms
branch
from
e8b8959 to
94e4dd3
Compare
…ic#2858) * init checkin * Update comment * Change user to actor * Resolve TODOs
What does this PR do?
This PR adds developer documentation illustrating the Fleet-managed Agent upgrade process, focussing on the communications amongst the various components involved.
Why is it important?
For Agent and Fleet developers to understand how Agent upgrades work.
Review suggestions
View the rich diff of the file.