-
Story
-
Resolution: Unresolved
-
Undefined
-
None
-
None
-
13
-
False
-
None
-
False
-
SECFLOWOTL-222 - Argo Agent: Sync protocol enhancements to fix known failure cases
-
-
Story (Required)
Background and Approach (Required)
See Challenging Argo CD Agent Cases v2 document (https://docs.google.com/document/d/1NwQaYxmRva8irx_clPTqRu_YQ8MNeFNXSTJoOgv7ZRs/edit?tab=t.0) for details on new messages and behaviour changes.
This new version of the document:
- Uses checksums to reduce bandwidth, which should significantly reduce bandwidth usage during agent/principal startup
- The previous version (v1) did not have a mechanism for the agent to periodically resync individual entities (e.g. applications). We now have ‘request-update’ for this. In v1, we instead had the ‘hello’ message, which refreshed all applications all at once (which would not scale as well in large environments).
- Better names for messages, to make it more explicit that they are used to synchronized entities
I was hoping to be able to split this into multiple stories, but since all 4 messages interact with each other, they need to be implemented as a single atomic unit.
Acceptance Criteria (Mandatory)
- New messages and message behaviour implemented in argocd-agent, as described in document.
- Unit tests
- Existing E2E tests still passing
- We should implement new E2E tests as part of this, BUT this story is already large enough as it is. So I will instead open a separate story for those.
- depends on
-
GITOPS-5667 Argo agent: Agents should keeping trying to send events to the principal
- Review
- is depended on by
-
GITOPS-5921 Argo CD Agent: Implement E2E test framework updates and E2E tests for 'challenging cases' A and B
- New
-
GITOPS-5922 Argo CD Agent: Investigate feasability of E2E framework updates for 'challenging cases' C, D, and E
- New