feat: add source sync observability#103
Conversation
|
Droid encountered an error —— View job Droid is reviewing code and running a security check… |
b5cde13 to
3afaf62
Compare
|
Droid finished @jonathanhaaswriter's task —— View job Validated 4 candidates and approved all 4. The source-level observability work is valuable, but the PR still has real regressions around rate limiting, wake-triggered sync behavior and error reporting, and dialog availability on unhealthy Google Workspace connectors. |
3afaf62 to
894894f
Compare
|
Droid finished @jonathanhaaswriter's task —— View job Validated 3 candidates and approved all 3. The new source observability surface still has real correctness gaps: per-source Sync ignores the selected stream, queued backfills look successful before workers run, and failed Directory or OAuth sweeps can appear current even when the underlying data is stale. |
894894f to
2584fe6
Compare
|
Droid finished @jonathanhaaswriter's task —— View job Validated all 3 review candidates for PR #103 and wrote Approved and posted 3 batched review comments:
|
2584fe6 to
655fe79
Compare
|
Droid finished @jonathanhaaswriter's task —— View job Validated 4 candidates and approved all 4. The source-level observability work is valuable, but the PR still has actionable correctness gaps around dialog state races, durable backfill queuing, wake-listener error handling, and BigQuery queue attribution. |
|
Droid finished @jonathanhaaswriter's task —— View job Validated 1 candidate and approved 1.
|
|
Droid finished @jonathanhaaswriter's task —— View job Validated all 2 review candidates for PR #103 and wrote Approved and posted 2 batched review comments:
Overall assessment: the source-level observability work is useful, but the PR still hides real wake-triggered setup failures for Reports and BigQuery streams. |
|
Droid finished @jonathanhaaswriter's task —— View job Validated all 3 review candidates for PR #103 and wrote Approved and posted 3 batched review comments:
Overall assessment: the source-level observability work is useful, but the PR still has actionable correctness gaps in once-mode wake draining, cross-connector Sync all state, and errored source row-count reporting. |
|
Droid finished @jonathanhaaswriter's task —— View job Validated all 2 review candidates for PR #103 and wrote Approved and posted 2 batched review comments:
Overall assessment: The source observability work is useful, but the PR still has two actionable correctness gaps in concurrent row actions and once-mode OAuth wake handling. |
|
Droid finished @jonathanhaaswriter's task —— View job Validated 1 candidate and approved it. The new source-level sync UI still exposes a real correctness gap: OAuth sync can be manually queued before Directory has seeded identities, which makes Shadow IT look current without scanning any users. |
|
Droid finished @jonathanhaaswriter's task —— View job I validated the lone remaining candidate and rejected it, so no new review comments were posted. The claimed same-row double-click race is not a high-confidence bug because React flushes the busy-state update between discrete clicks, and existing review feedback already covered the real busy-state regression this patch fixed. |
Summary
Validation
Note: make test-go-db could not run locally because Docker/Colima is not running (missing /Users/jonathan/.colima/default/docker.sock).