Add OTEP #4947 OTEL-compatible context storage by jbachorik · Pull Request #347 · DataDog/java-profiler

jbachorik · 2026-01-29T16:48:17Z

What does this PR do?:

Exposes trace context via an OTEP #4947-compliant TLS pointer (otel_thread_ctx_v1),
discoverable by external profilers through the ELF dynsym table. OTEL context storage
is the sole context storage mode; there is no legacy profiler-only mode.

Each thread gets a 640-byte OtelThreadContextRecord containing:

W3C trace_id (128-bit) and span_id (64-bit) in big-endian byte encoding
Custom attributes encoded as key-index + length + UTF-8 value pairs in attrs_data[612]

A ContextApi class provides signal-safe read and snapshot operations on the OTEP record.

Context data flow

flowchart LR
    T["Java tracer"]

    subgraph thread ["Per-thread (ProfiledThread)"]
        REC["OtelThreadContextRecord\nvalid · trace_id · span_id · attrs_data"]
        SC["JFR sidecar\nu32 encodings · LRS"]
    end

    T -- "put(): valid=0 → write → valid=1" --> REC
    T -- "put(): LRS\nsetContextAttribute(): u32 encoding" --> SC
    T -- "setContextAttribute(): UTF-8 value" --> REC

    REC -- "otel_thread_ctx_v1 (dynsym)" --> EP["External profiler\n(OTEP #4947)"]
    SC -- "O(1), signal-safe" --> SH["Signal handler\n→ JFR writer"]
    REC -- "valid + span_id" --> SH

Publication protocol (detach/attach)

sequenceDiagram
    participant W as Java tracer (writer)
    participant rec as OtelThreadContextRecord
    participant R as Reader (signal handler / external profiler)

    Note over W,R: Thread init: otel_thread_ctx_v1 permanently points to record

    W->>rec: valid = 0
    W->>rec: storeFence()
    Note over R: load valid → 0, skip record
    W->>rec: write trace_id, span_id, attrs_data, LRS
    W->>rec: storeFence()
    W->>rec: valid = 1
    Note over R: load valid → 1, read record safely
    R->>rec: read trace_id, span_id, attrs_data

Context setting API:

New: ThreadContext.put(localRootSpanId, spanId, traceIdHigh, traceIdLow) — writes
trace_id + span_id to the OTEP record, stores localRootSpanId in ProfiledThread for
O(1) signal-safe reads (used for endpoint correlation in JFR).
Existing (deprecated): ThreadContext.put(spanId, rootSpanId) — delegates to the
4-arg overload with traceIdHigh=0 and traceIdLow=spanId. Marked @Deprecated; callers
needing correct OTEP interop must switch to the 4-arg API.
New: ThreadContext.setContextAttribute(keyIndex, value) — writes custom attributes.
Encodes the string value into both the sidecar (u32 Dictionary encoding for JFR) and
attrs_data (UTF-8 for external profilers) via ByteBuffer. JNI is called only on cache
miss to register a new constant.

JFR tag encoding sidecar:

ProfiledThread caches pre-computed Dictionary encodings (_otel_tag_encodings[]) and
_otel_local_root_span_id alongside the OTEP record. The ByteBuffer write path populates
both the OTEP attrs_data (raw strings for external profilers) and the sidecar (u32
encodings for our JFR writer). Signal handlers read only the sidecar — O(1) array index,
zero string lookups or hash-table probes per sample.

Motivation:

Enable external profilers (e.g. OTel eBPF profiler) to discover and read the Java
profiler's trace context using the standardized OTEP #4947 protocol, while preserving
localRootSpanId for Datadog endpoint correlation.

Additional Notes:

Signal safety: attribute string values are pre-registered in the Dictionary from the
JNI thread; the signal handler reads only the sidecar (no Dictionary access at all)
Publication protocol: 3-step with release barriers — set valid=0 + storeFence, write
trace/span/attrs data, storeFence + set valid=1 — ensuring signal handlers never see
partially-written records. The TLS pointer otel_thread_ctx_v1 is set permanently on
thread init and never modified during context writes.
Thread lifecycle: OtelThreadContextRecord is zeroed and the TLS pointer nulled in
releaseFromBuffer() to prevent external profilers from dereferencing a recycled record
Sidecar fields are zeroed in releaseFromBuffer() before the OTEL record is cleared
Architecture docs: doc/architecture/TLSContext.md

How to test the change?:

Automated tests (7 tests):

./gradlew :ddprof-test:testRelease --tests "com.datadoghq.profiler.context.OtelContextStorageModeTest"

Covers: OTEL context round-trip with 128-bit trace ID + localRootSpanId, custom
attributes, attribute overflow, sequential context updates, thread isolation,
span transition clearing attributes, attribute cache isolation.

Symbol verification:

nm -D libjavaProfiler.so | grep otel_thread_ctx_v1

For Datadog employees:

If this PR touches code that signs or publishes builds or packages, or handles
credentials of any kind, I've requested a review from @DataDog/security-design-and-guidance.
This PR doesn't touch any of that.
JIRA: PROF-14224

ivoanjo

Gave it a pass! There were a few of "uuuuh what? 👀 moments". I guess AI?

ddprof-lib/src/main/cpp/arguments.cpp

ddprof-lib/src/main/cpp/context_api.cpp

ddprof-lib/src/main/cpp/flightRecorder.cpp

ddprof-lib/src/main/cpp/javaApi.cpp

ddprof-lib/src/main/java/com/datadoghq/profiler/OTelContext.java

ddprof-test/src/test/java/com/datadoghq/profiler/context/OtelContextStorageModeTest.java

ddprof-test/src/test/java/com/datadoghq/profiler/context/ProcessContextTest.java

doc/architecture/OtelContextStorage.md

doc/OTelContextReference.md

jbachorik · 2026-01-30T12:42:16Z

Gave it a pass! There were a few of "uuuuh what? 👀 moments". I guess AI?

Yes. Ah, you shouldn't have spent a lot of time on this ... this is just a dirty AI generated scaffolding, hence still draft ..
But, since you already did, thanks for the valuable input!

…of reference library **What does this PR do?** This PR updates the vendored version of the OTel process context support with the latest code from open-telemetry/sig-profiling#79 (which will soon be merged upstream). **Motivation:** This latest version of the library fixes a number of small issues, but more importantly includes support for the thread context configuration keys which we need for #347 (so we will be able to shed a bunch of code once we rebase on top of this one). **Additional Notes:** From the Java API consumer side, not a lot changes with this version. **How to test the change?** Validate tests still pass!

dd-octo-sts · 2026-03-19T07:59:22Z

CI Test Results

Run: #24146663568 | Commit: e6726b7 | Duration: 23m 17s (longest job)

✅ All 32 test jobs passed

Status Overview

JDK	glibc-aarch64/debug	glibc-amd64/debug	musl-aarch64/debug	musl-amd64/debug
8	-	✅	-	-
8-ibm	-	✅	-	-
8-j9	✅	✅	-	-
8-librca	-	-	✅	✅
8-orcl	-	✅	-	-
11	-	✅	-	-
11-j9	✅	✅	-	-
11-librca	-	-	✅	✅
17	✅	✅	-	-
17-graal	✅	✅	-	-
17-j9	✅	✅	-	-
17-librca	-	-	✅	✅
21	✅	✅	-	-
21-graal	✅	✅	-	-
21-librca	-	-	✅	✅
25	✅	✅	-	-
25-graal	✅	✅	-	-
25-librca	-	-	✅	✅

Legend: ✅ passed | ❌ failed | ⚪ skipped | 🚫 cancelled

Summary: Total: 32 | Passed: 32 | Failed: 0

Updated: 2026-04-08 16:56:50 UTC

pr-commenter · 2026-03-19T08:08:27Z

Integration Tests

✅ All 40 integration tests passed

📊 Dashboard · 👷 Pipeline · 📦 b69e944f

ivoanjo · 2026-03-20T09:05:17Z

ddprof-lib/src/main/cpp/arguments.h

+ * PROFILER: Use existing TLS-based storage (default, proven async-signal safe)
+ * OTEL: Use OTEP #4947 TLS pointer storage (discoverable by external profilers)


Ack! As I said above, I think having a "press button in case of emergency" makes sense, but overall I really think it's important to get this to a situation where OTel is the default in java-profiling and carries no downsides (even if that means changing the spec, as many times as needed).

ivoanjo

Here's my first pass of comments before lunch, be back soon!

ddprof-lib/src/main/cpp/flightRecorder.cpp

ddprof-lib/src/main/cpp/context_api.h

ddprof-lib/src/main/cpp/context_api.cpp

ivoanjo

Another round of comments, haven't gone through everything 😅 C++ is cool but hard :)

ddprof-lib/src/main/cpp/context_api.cpp

ddprof-lib/src/main/cpp/thread.h

ivoanjo · 2026-03-20T17:11:44Z

ddprof-lib/src/main/cpp/context_api.cpp

+#ifndef OTEL_PROCESS_CTX_NO_READ
+        // Re-publish the process context with thread_ctx_config.
+        // Guard: otel_process_ctx_read() is only available when the read API is compiled in.
+        // We need to read the current context and re-publish with the config
+        otel_process_ctx_read_result read_result = otel_process_ctx_read();
+        if (read_result.success) {
+            otel_process_ctx_data data = {
+                .deployment_environment_name = read_result.data.deployment_environment_name,
+                .service_instance_id = read_result.data.service_instance_id,
+                .service_name = read_result.data.service_name,
+                .service_version = read_result.data.service_version,
+                .telemetry_sdk_language = read_result.data.telemetry_sdk_language,
+                .telemetry_sdk_version = read_result.data.telemetry_sdk_version,
+                .telemetry_sdk_name = read_result.data.telemetry_sdk_name,
+                .resource_attributes = read_result.data.resource_attributes,
+                .extra_attributes = read_result.data.extra_attributes,
+                .thread_ctx_config = &config,
+            };
+
+            otel_process_ctx_publish(&data);
+            otel_process_ctx_read_drop(&read_result);
+        }
+#endif


Oh interesting, I'll admit I hadn't considered that it would be nice for the process context impl to have a "append thread context config to an existing process context".

So I guess this is the homegrown version of that -- read existing, append, write again.

Would it be useful for me to add a otel_process_ctx_update_thread_ctx to do this? In particular can be done without requiring the cost of the reader (e.g. we could compile with the reader disabled).

ddprof-lib/src/main/cpp/context_api.cpp

ddprof-lib/src/main/cpp/flightRecorder.cpp

ddprof-lib/src/main/cpp/javaApi.cpp

ivoanjo

Gave it a full pass! Really excited to see how this is evolving! Overall I think there's a bit of code that we can rip out/simplify/improve :D

ddprof-lib/src/main/cpp/context.cpp

ddprof-lib/src/main/cpp/flightRecorder.cpp

ddprof-lib/src/main/cpp/context_api.cpp

ddprof-lib/src/main/cpp/otel_context.cpp

ddprof-test/src/test/java/com/datadoghq/profiler/context/OtelContextStorageModeTest.java

ddprof-test/src/test/java/com/datadoghq/profiler/ContextSanityTest.java

doc/architecture/OtelContextStorage.md

doc/architecture/OtelThreadContextSharing.md

ivoanjo

Gave it another pass! :D

ddprof-lib/src/main/cpp/thread.h

ddprof-lib/src/main/cpp/thread.cpp

ddprof-lib/src/main/java/com/datadoghq/profiler/BufferWriter.java

ddprof-lib/src/main/cpp/javaApi.cpp

ddprof-lib/src/main/java/com/datadoghq/profiler/ThreadContext.java

ddprof-test/src/test/java/com/datadoghq/profiler/context/OtelContextStorageModeTest.java

ddprof-lib/src/main/cpp/otel_process_ctx.cpp

jbachorik · 2026-03-25T15:06:24Z

Reply to #347 (comment): Done, added // For testing only — production code gets context from the tracer, not the profiler. right above the function.

- Remove OtelContexts::u64ToBytes (never called) - Remove ProfiledThread::setOtelTagEncoding, setOtelLocalRootSpanId (Java writes sidecar directly via ByteBuffer) - Eliminate ContextApi::_attribute_keys member array, freeAttributeKeys, and shutdown — otel_process_ctx_publish copies strings, no keepalive needed - registerAttributeKeys0: always publish even with count==0 so datadog.local_root_span_id (index 0) is always in the process context - Remove stale tlsPtrBuffer comment (PR-evolution artifact) - Add ThreadContext.readTraceId() for test-side trace ID verification Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

- setContextDirect non-clear path now zeros sidecar tag encodings and trims attrs_data to LRS_ENTRY_SIZE so span A's attributes don't leak into span B when setContext is called without an intervening clear - Add testSpanTransitionClearsAttributes to cover this exact path - Use readTraceId() in testOtelStorageModeContext for trace ID round-trip - Fix TLSContext.md write order (sidecar write was shown before trace ID writes; code does it after; also add the new attr-reset step) Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

- Remove stale lrsCache reference (LRS uses inline hex encoding, no cache) - Fix LRS encoding description: hex (16-char zero-padded), not decimal; clarify its fixed position at attrs_data[0..17] - Expand step 3 description to mention the fixed LRS entry is retained when resetting attrs_data_size to LRS_ENTRY_SIZE Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

…st comment Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

…ebug Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

HEX_DIGITS is byte[]; sb.append(byte) widens to append(int) and appends the decimal ASCII code (e.g. '5'=53 -> "53") instead of the character. The JIT happens to mask this in release builds but the interpreter exposes it in debug builds. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

- clearContextAttribute: plain putInt inside detach/attach window - setContextAttributeDirect: same; dict-full path reuses clearContextAttribute - Remove redundant testRepeatedContextWrites and testNestedContextUpdates - TLSContext.md: fix CPU store buffer claim (compiler reordering only for same-thread); update code snippet and test list Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

- thread.h/wallClock.cpp: guard _otel_local_root_span_id update behind context_valid; ContextApi::get returns valid=0 between detach/attach causing the sidecar to be clobbered with 0 before attach completes - JavaProfiler.java/javaApi.cpp: remove dead registerConstant() and its JNI function (no callers since ContextSetter was removed) - otel_process_ctx.cpp: remove duplicate #include <sys/syscall.h> - ThreadContext.java: fix detach() javadoc and static cache comment - TLSContext.md: clarify ~35 ns = spanLifecycle + clearContext Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

valid stays 0 after clear (no attach()), so no reader can observe attrs_data. The next setContext call resets attrsDataSize before attach() anyway. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

Static cache arrays allowed a thread to read a stale encoding from a slot evicted by another thread. Per-thread (instance) arrays mean only the owning thread ever accesses them — no races, no barriers needed. Drop fullFence from the hit path and storeFence from the write path as a result. Each thread pays one JNI registerConstant0() call per unique value on first encounter; encodings are stable for the JVM lifetime. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

"FB" and "Ea" share hashCode 2236 (slot 188). Test verifies that two threads writing those values simultaneously each read back their own correct value — impossible with the old static cache. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

@jbachorik

- Rename initializeOtelTls0/isOtelContextInitialized/markOtelContextInitialized to drop Otel prefix (only one impl exists now) - Factor clearContextDirect() out of setContextDirect() - Use Integer.BYTES instead of magic 4 for sidecar byte offsets - Drop dead writeAndReleaseLong/writeAndReleaseInt from BufferWriter - Add LOG_WARN when registerAttributeKeys keys exceed capacity - Clarify comments: musl TLS deadlock, contextValid vs zero spans, recycling vs valid-flag, plain put safety in attach() - Fix javadocs: clearContext, setContext, readContextAttribute, readTraceId Rationale: PR review by @jbachorik

…lFence - Rename custom_labels_current_set_v2 -> otel_thread_ctx_v1 per OTEP #4947 final naming - Remove fullFence() (storeFence is sufficient; no load ordering needed) - Make bytesToU64 file-local static; drop BytesToU64 from header - Use plain putLong for LRS sidecar (no cross-thread ordering needed) - Consolidate testMaxValueContext and testOtelModeClearContext into testSequentialContextUpdates

- Inline getSpanId into ContextApi::get(); remove OtelContexts class - Remove bytesToU64 and cassert (now unused in otel_context.cpp) - Fix TLSContext.md: scheduler rq_lock barrier makes stores visible, not "thread paused"; DMB ISHST is required on ARM, not a side effect - Fix otel_context.h: pointer is nulled on thread exit, not "never nulled" - Fix TLSContext.md: pointer is nulled on thread exit to prevent use-after-recycle; external profilers must null-check before deref

- Replace custom_labels_current_set_v2 with otel_thread_ctx_v1 in diagram - Remove deleted OtelContexts/getSpanId/bytesToU64 references - Update writeOrderedLong -> putLong in diagrams and protocol timeline - Fix process-wide -> per-thread cache label - Remove phantom storeFence from attr-cache pseudo-code (cache is per-thread) - Remove phantom putLongVolatile/setVolatile from BufferWriter diagram - Update ContextApi::get() snippet to match actual inlined implementation - Add missing numAttrs declaration in writeCurrentContext snippet

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

…view Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

@param

- Add @param context_valid to lookupWallclockCallTraceId doxygen - Add static_assert linking OtelThreadContextRecord size to OTEL_MAX_RECORD_SIZE - Deprecate 2-arg JavaProfiler.setContext; clarify rootSpanId→localRootSpanId mapping - Explain return 0 in deprecated put(long,long) is for ABI compatibility - Fix "per-process" → "per-thread" encoding cache in TLSContext.md - Clarify x86 storeFence applies to Java 9+ VarHandle path in TLSContext.md - Suppress deprecation warnings at 3 legacy setContext call sites Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

- wallClock.h: replace assert with null guard for ProfiledThread race - thread.cpp: copy tag encodings only when ContextApi::get() returns true - ThreadContext: read spanId/rootSpanId directly from buffer, drop getContext0() JNI - jfrMetadata.cpp: document signal-safety precondition on reset() - ThreadContext: clarify 2-arg put() param mapping, note otepKeyIndex != 0 invariant - OtelContextStorageModeTest: add 1000-iteration stress test

- ThreadContextBenchmark: replace removed getContext() with getSpanId() - thread.cpp: use __atomic_store_n(RELEASE) for _otel_ctx_initialized - thread.h: document naturally-aligned u64 atomicity assumption - ThreadContext: document clearContextDirect valid=0 OTEP reader trade-off - context_api.h: move LOCAL_ROOT_SPAN_ATTR_INDEX to context_api.cpp (file-local) - perf report: remove stale getContext results, note replacement by direct buffer reads

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

jbachorik · 2026-04-08T17:15:18Z

@ivoanjo Thanks for the thorough review!

jbachorik added the AI label Jan 29, 2026

ivoanjo reviewed Jan 30, 2026

View reviewed changes

jbachorik changed the title ~~OTEL context conformance with tlsdesc_v1_dev spec~~ WIP: OTEL context conformance with tlsdesc_v1_dev spec Feb 2, 2026

ivoanjo mentioned this pull request Mar 11, 2026

[PROF-12646] Update OTel process context support with latest version of reference library #414

Merged

3 tasks

jbachorik force-pushed the jb/otel_ctx branch from 4850579 to 20b9f85 Compare March 19, 2026 07:46

jbachorik force-pushed the jb/otel_ctx branch from 20b9f85 to 351f2b5 Compare March 19, 2026 17:31

jbachorik changed the title ~~WIP: OTEL context conformance with tlsdesc_v1_dev spec~~ Add OTEP #4947 context storage with custom attributes Mar 19, 2026

jbachorik changed the title ~~Add OTEP #4947 context storage with custom attributes~~ Add OTEP #4947 OTEL-compatible context storage Mar 19, 2026

jbachorik force-pushed the jb/otel_ctx branch 4 times, most recently from 8849603 to 9bbf480 Compare March 19, 2026 20:26

ivoanjo reviewed Mar 20, 2026

View reviewed changes

jbachorik force-pushed the jb/otel_ctx branch 2 times, most recently from 88a03d5 to 73f90c9 Compare March 20, 2026 10:10

jbachorik marked this pull request as ready for review March 20, 2026 10:46

jbachorik requested a review from a team as a code owner March 20, 2026 10:46

jbachorik added the AI: Author Signed-off label Mar 20, 2026

jbachorik force-pushed the jb/otel_ctx branch from 73f90c9 to 4243a48 Compare March 20, 2026 11:49

ivoanjo reviewed Mar 20, 2026

View reviewed changes

jbachorik force-pushed the jb/otel_ctx branch from e2510f3 to 349ebbc Compare March 21, 2026 19:57

ivoanjo reviewed Mar 24, 2026

View reviewed changes

jbachorik force-pushed the jb/otel_ctx branch from c10d22e to fea2bb5 Compare March 24, 2026 19:31

ivoanjo reviewed Mar 25, 2026

View reviewed changes

jbachorik and others added 25 commits April 8, 2026 18:30

writeOrderedLong/writeAndReleaseLong: drop wasteful read-back; fix te…

ee400c4

…st comment Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

Fix testOtelStorageModeContext: stop profiler before readTraceId in d…

d7d21b6

…ebug Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

Drop redundant attrsDataSize reset in clearContext path

9b14057

valid stays 0 after clear (no attach()), so no reader can observe attrs_data. The next setContext call resets attrsDataSize before attach() anyway. Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

Warn against high-cardinality values in setContextAttribute javadoc

74f3e99

Co-Authored-By: Claude Sonnet 4.6 (1M context) <noreply@anthropic.com>

fix: remove duplicate destructor in ProfiledThread (rebase artifact)

24d8781

fix: remove stale fullFence javadoc reference in BufferWriter

3236b71

fix(test): remove deleted writeVolatile* calls from BufferWriterTest

8555d61

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

fix: address dead code, stale test refs, and doc imprecisions from re…

7a11ef2

…view Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

docs: update benchmark report with 2026-04-07 results

236dbf8

test: add process context preservation test for registerAttributeKeys

fe5aef1

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

jbachorik force-pushed the jb/otel_ctx branch from b26369c to fe5aef1 Compare April 8, 2026 16:30

jbachorik merged commit 2c49bab into main Apr 8, 2026
98 checks passed

jbachorik deleted the jb/otel_ctx branch April 8, 2026 17:08

github-actions bot added this to the 1.41.0 milestone Apr 8, 2026

		* PROFILER: Use existing TLS-based storage (default, proven async-signal safe)
		* OTEL: Use OTEP #4947 TLS pointer storage (discoverable by external profilers)

Conversation

jbachorik commented Jan 29, 2026 • edited by atlassian bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

ivoanjo left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jbachorik commented Jan 30, 2026

Uh oh!

dd-octo-sts bot commented Mar 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

CI Test Results

Status Overview

Uh oh!

pr-commenter bot commented Mar 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Integration Tests

Uh oh!

ivoanjo Mar 20, 2026

Choose a reason for hiding this comment

Uh oh!

ivoanjo left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ivoanjo left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ivoanjo Mar 20, 2026

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ivoanjo left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

ivoanjo left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

jbachorik commented Jan 29, 2026 •

edited by atlassian bot

Loading

dd-octo-sts bot commented Mar 19, 2026 •

edited

Loading

pr-commenter bot commented Mar 19, 2026 •

edited

Loading