fix(chipingress): empty PublishBatch results with nil error counts as delivered

[cawthorne]

Summary

A nil-error PublishBatch response carrying no per-event results is reported to every event's callback as ErrCodeResultsMismatch since dc8a1dc removed the len(resp.Results) > 0 guard in sendBatch. Servers that accept a batch without per-event detail do exist: the local-cre chip-router v1.x forwards every event to Kafka but returns an empty results array.

With the durable emitter on top, this is not a cosmetic mismatch:

• the durable emitter treats every event as undelivered and retransmits after 60s — forever, because the ack never comes;
• the node re-sends its whole event stream endlessly;
• the Kafka topic fills with duplicate retransmits and consumers drown.

Evidence (chainlink v2.67.2-rc.0 node vs chip-router v1.0.1)

Clean A/B against the same router, same environment, only the chainlink-common pin differing (v0.11.2-0.20260914191328 (chipingress v0.0.11-0.20260724142814) vs v0.11.2-0.20260929093916 (chipingress v0.0.11-0.20260915184316)):

	old pin (guard present)	new pin (guard removed)
DurableEmitter: failed to deliver event. Relying on retransmit.	0	301,385 on one node, first at 12s after boot, zero successful acks
client error	—	-1: server returned 0 results for 1/500 events
Kafka topic rate	modest	~4,000 msg/s (duplicate retransmits; counters inflated ~20x)
metric-event delivery latency	~seconds	gaps of 1.5–4.5 min (buffer backup)

The events were delivered to Kafka each time (the router forwards on receive) — only the acks were missing — so the retransmit storm duplicates the entire stream rather than losing data. The delivery-latency gaps are long enough to break downstream freshness-gated consumers (e.g. recording rules gating on time() - <metric>_ts < 120).

Fix

Restores the pre-dc8a1dc3 semantics in sendBatch:

• nil error + empty results → every callback resolves with nil (delivered);
• nil error + non-empty results → still correlated per event (unchanged);
• partial results → still RESULTS_MISMATCH for the missing tail (unchanged, existing test covers it).

Adds a regression test for the zero-results case.

Alternatives considered

• Router side: chip-router populates per-event results — also valid, but any deployed router without them turns every producer into an eternal retransmit loop, so the client tolerance is the robust side to fix.
• Config knob: Telemetry.ChipIngressBatchEmitterEnabled = false falls back to the legacy per-event emitter and avoids the contract — but it's a throughput regression at scale and the field is framework-generated in CRE environments, so it isn't reachable for affected deployments.

[github-actions]

👋 cawthorne, thanks for creating this pull request!

To help reviewers, please consider creating future PRs as drafts first. This allows you to self-review and make any final changes before notifying the team.

Once you're ready, you can mark it as "Ready for review" to request feedback. Thanks!

[github-actions]

📊 API Diff Results

No changes detected for module github.com/smartcontractkit/chainlink-common/pkg/chipingress

View full report

[Copilot]

Copilot review overview

🟢 Approval recommended

The focused compatibility fix matches the stated semantics and includes appropriate regression coverage.

Review effort: Balanced
Findings: None

What changed in this PR

Restores compatibility with batch servers that acknowledge delivery without per-event results, preventing unnecessary durable-emitter retransmissions.

Changes:

• Treats successful empty PublishBatch responses as delivered.
• Adds regression coverage while preserving partial-result mismatch handling.

File	Description
pkg/​chipingress/​batch/​client.go	Handles empty successful responses as batch success.
pkg/​chipingress/​batch/​client_test.go	Tests callbacks for empty-result responses.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[cawthorne]

@pkcll

It seems this might have been intentional:
https://ainv.smartcontract.com/findings/79362753-0fef-4a50-934c-572e8f7c504d

This might actually be the fix:
smartcontractkit/chainlink-testing-framework#2861

c.c. @jmank88