[CI] Push native debug symbols to Datadog Symbolication server by gleocadie · Pull Request #6081 · DataDog/dd-trace-dotnet

gleocadie · 2024-09-25T13:16:10Z

Summary of changes

Pushing native debug symbols to datadog infra.

Reason for change

Reports sent by the crash tracker (when application is crashing) contains unresolved callstacks. When every pieces of the puzzle will be in place, the backend will be able to resolve the instruction pointers and by pushing our native debug symbols to datadog infra, we make sure that instruction pointers related to our native binaries will be correctly resolved.

Implementation details

Create a github action in charge of setting every things we need to send the symbols to the backend (setup datadog-ci and call it)
Add a manual workflow, so we can push old version
Update create release workflow to use the actions

Test coverage

Run multiple jobs successfully https://github.com/DataDog/dd-trace-dotnet/actions/runs/11035801336

Other details

datadog-ddstaging · 2024-09-25T13:38:20Z

Datadog Report

Branch report: andrew/symbols-gitlab-job
Commit report: bdbc161
Test service: dd-trace-dotnet

✅ 0 Failed, 150169 Passed, 448 Skipped, 1h 58m 59.9s Total Time

andrewlock · 2024-09-25T13:54:30Z

Execution-Time Benchmarks Report ⏱️

Execution-time results for samples comparing the following branches/commits:

Execution-time benchmarks measure the whole time it takes to execute a program. And are intended to measure the one-off costs. Cases where the execution time results for the PR are worse than latest master results are shown in red. The following thresholds were used for comparing the execution times:

Welch test with statistical test for significance of 5%
Only results indicating a difference greater than 5% and 5 ms are considered.

Note that these results are based on a single point-in-time result for each branch. For full results, see the dashboard.

Graphs show the p99 interval based on the mean and StdDev of the test run, as well as the mean value of the run (shown as a diamond below the graph).

gantt
    title Execution time (ms) FakeDbCommand (.NET Framework 4.6.2) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6081) - mean (71ms)  : 68, 73
     .   : milestone, 71,
    master - mean (71ms)  : 68, 73
     .   : milestone, 71,

    section CallTarget+Inlining+NGEN
    This PR (6081) - mean (1,107ms)  : 1087, 1126
     .   : milestone, 1107,
    master - mean (1,110ms)  : 1087, 1132
     .   : milestone, 1110,

gantt
    title Execution time (ms) FakeDbCommand (.NET Core 3.1) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6081) - mean (110ms)  : 106, 113
     .   : milestone, 110,
    master - mean (110ms)  : 107, 112
     .   : milestone, 110,

    section CallTarget+Inlining+NGEN
    This PR (6081) - mean (772ms)  : 755, 789
     .   : milestone, 772,
    master - mean (775ms)  : 759, 791
     .   : milestone, 775,

gantt
    title Execution time (ms) FakeDbCommand (.NET 6) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6081) - mean (93ms)  : 90, 96
     .   : milestone, 93,
    master - mean (94ms)  : 90, 97
     .   : milestone, 94,

    section CallTarget+Inlining+NGEN
    This PR (6081) - mean (730ms)  : 714, 746
     .   : milestone, 730,
    master - mean (734ms)  : 713, 756
     .   : milestone, 734,

gantt
    title Execution time (ms) HttpMessageHandler (.NET Framework 4.6.2) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6081) - mean (191ms)  : 188, 194
     .   : milestone, 191,
    master - mean (191ms)  : 188, 194
     .   : milestone, 191,

    section CallTarget+Inlining+NGEN
    This PR (6081) - mean (1,205ms)  : 1182, 1227
     .   : milestone, 1205,
    master - mean (1,199ms)  : 1176, 1222
     .   : milestone, 1199,

gantt
    title Execution time (ms) HttpMessageHandler (.NET Core 3.1) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6081) - mean (278ms)  : 272, 283
     .   : milestone, 278,
    master - mean (277ms)  : 272, 282
     .   : milestone, 277,

    section CallTarget+Inlining+NGEN
    This PR (6081) - mean (942ms)  : 922, 963
     .   : milestone, 942,
    master - mean (941ms)  : 920, 962
     .   : milestone, 941,

gantt
    title Execution time (ms) HttpMessageHandler (.NET 6) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (6081) - mean (266ms)  : 262, 271
     .   : milestone, 266,
    master - mean (266ms)  : 263, 270
     .   : milestone, 266,

    section CallTarget+Inlining+NGEN
    This PR (6081) - mean (930ms)  : 906, 955
     .   : milestone, 930,
    master - mean (930ms)  : 910, 951
     .   : milestone, 930,

andrewlock

LGTM! Just a couple of small questions about verifying that the action has succeeded.

Once we've established that this does describe failures correctly and/or we know how to verify it worked, we should add a note about this to the release notes (just after the "run the create_drafe_release" section) to describe what happens, and how to remediate (running the workflow manually)

andrewlock · 2024-09-26T08:45:14Z

+      - name: Prepare debug symbols for publishing
+        shell: bash
+        run: |
+          mkdir debug_symbols
+          tar zxcf linux-native-symbols.tar.gz -C debug_symbols
+
+      - name: 'Push debug symbols to datadog'
+        uses: ./.github/actions/publish-debug-symbols
+        with:
+          symbols_folder: "debug_symbols"
+          preprod_key: "${{ secrets.DD_PREPROD_API_KEY }}"
+          prod_key: "${{ secrets.DD_API_KEY }}"


Question: is it better to do this before or after we publish to NuGet/create the release? 🤔

If before, then if there's any issues the good news is we can correct them first, and we can't create releases without symbols. The bad news is it would block the release if there's a problem, or if datadog is down or something 🤔

I think I'm inclined to go with your approach, especially as we can also do this manually using the other action if there is a problem later 🤔

andrewlock · 2024-09-26T08:45:45Z

+  workflow_dispatch:
+    inputs:
+      version:
+        description: 'An already released version of .NET APM'


Suggested change

description: 'An already released version of .NET APM'

description: 'An already released version of .NET APM, without the "v" suffix, e.g. "3.3.1"'

andrewlock · 2024-09-26T08:48:37Z

+      run: |
+        if [ -n "${{ inputs.prod_key }}" ]; then
+          echo "Push symbols to prod env"
+          DATADOG_API_KEY="${{ inputs.prod_key }}" DD_BETA_COMMANDS_ENABLED=1 datadog-ci elf-symbols upload ${{inputs.symbols_folder}}


If this fails, does it fail the action? We should try to test/make sure that it does, so that we know if/when this fails (and can remediate by running the action manually).

Also, is there any way for us to verify inside datadog that these have been uploaded correctly, whether that's automatic or manual?

yes if it fails, the action/job is marked as failed https://github.com/DataDog/dd-trace-dotnet/actions/runs/11052646040

Also, is there any way for us to verify inside datadog that these have been uploaded correctly, whether that's automatic or manual?

Not that I know of. I have been told that if datadog-ci succeeds the symbols will be there ~5-10min.
Besides, It requires specific rights to query the service (AWS bucket) and not sure it's possible outside of Datadog. I'll see with Nicolas Savoir on that topic.

gleocadie force-pushed the andrew/symbols-gitlab-job branch 15 times, most recently from bdbc161 to cdde660 Compare September 25, 2024 15:28

github-actions Bot added the area:builds project files, build scripts, pipelines, versioning, releases, packages label Sep 25, 2024

Push linux debug symbols to datadog symbolication server

6a85a6c

gleocadie force-pushed the andrew/symbols-gitlab-job branch from cdde660 to 6a85a6c Compare September 25, 2024 15:34

gleocadie changed the title ~~Andrew/symbols gitlab job~~ [CI] Push native debug symbols to Datadog Symbolication server Sep 25, 2024

andrewlock approved these changes Sep 26, 2024

View reviewed changes

gleocadie marked this pull request as ready for review September 26, 2024 10:33

gleocadie requested a review from a team as a code owner September 26, 2024 10:33

Address PR comment

f1d5612

gleocadie force-pushed the andrew/symbols-gitlab-job branch from f6b394f to f1d5612 Compare September 26, 2024 13:02

andrewlock approved these changes Sep 26, 2024

View reviewed changes

gleocadie merged commit f64cde1 into master Sep 26, 2024

gleocadie deleted the andrew/symbols-gitlab-job branch September 26, 2024 13:59

github-actions Bot added this to the vNext-v3 milestone Sep 26, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

[CI] Push native debug symbols to Datadog Symbolication server#6081

[CI] Push native debug symbols to Datadog Symbolication server#6081
gleocadie merged 2 commits into
masterfrom
andrew/symbols-gitlab-job

gleocadie commented Sep 25, 2024 •

edited

Loading

Uh oh!

datadog-ddstaging Bot commented Sep 25, 2024 •

edited

Loading

Uh oh!

andrewlock commented Sep 25, 2024 •

edited

Loading

Uh oh!

andrewlock left a comment

Uh oh!

andrewlock Sep 26, 2024

Uh oh!

gleocadie Sep 26, 2024

Uh oh!

andrewlock Sep 26, 2024

Uh oh!

gleocadie Sep 26, 2024

Uh oh!

andrewlock Sep 26, 2024

Uh oh!

gleocadie Sep 26, 2024 •

edited

Loading

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

	description: 'An already released version of .NET APM'
	description: 'An already released version of .NET APM, without the "v" suffix, e.g. "3.3.1"'

Uh oh!

Conversation

gleocadie commented Sep 25, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary of changes

Reason for change

Implementation details

Test coverage

Other details

Uh oh!

datadog-ddstaging Bot commented Sep 25, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Datadog Report

Uh oh!

andrewlock commented Sep 25, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Execution-Time Benchmarks Report ⏱️

Uh oh!

andrewlock left a comment

Choose a reason for hiding this comment

Uh oh!

andrewlock Sep 26, 2024

Choose a reason for hiding this comment

Uh oh!

gleocadie Sep 26, 2024

Choose a reason for hiding this comment

Uh oh!

andrewlock Sep 26, 2024

Choose a reason for hiding this comment

Uh oh!

gleocadie Sep 26, 2024

Choose a reason for hiding this comment

Uh oh!

andrewlock Sep 26, 2024

Choose a reason for hiding this comment

Uh oh!

gleocadie Sep 26, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

gleocadie commented Sep 25, 2024 •

edited

Loading

datadog-ddstaging Bot commented Sep 25, 2024 •

edited

Loading

andrewlock commented Sep 25, 2024 •

edited

Loading

gleocadie Sep 26, 2024 •

edited

Loading