run-e2e

name: run-e2e description: Run SkyWalking E2E tests locally disable-model-invocation: true argument-hint: "[test-case-path]"

Run SkyWalking E2E Test

Run an E2E test case using skywalking-infra-e2e. The user provides a test case path (e.g., simple/jdk, storage/banyandb, alarm).

Prerequisites

All tools require Go installed. Check .github/workflows/ for the exact e2e commit used in CI.

e2e CLI

Built from apache/skywalking-infra-e2e, pinned by commit in CI:

# Install the pinned commit
go install github.com/apache/skywalking-infra-e2e/cmd/e2e@<commit-id>

# Or clone and build locally (useful when debugging the e2e tool itself)
git clone https://github.com/apache/skywalking-infra-e2e.git
cd skywalking-infra-e2e
git checkout <commit-id>
make build
# binary is in bin/e2e — add to PATH or copy to $GOPATH/bin

swctl, yq, and other tools

E2E test cases run pre-install steps (see setup.steps in each e2e.yaml) that install tools into /tmp/skywalking-infra-e2e/bin. When running locally, you need these tools on your PATH.

swctl — SkyWalking CLI, used in verify cases to query OAP's GraphQL API. Pinned at SW_CTL_COMMIT in test/e2e-v2/script/env:

# Option 1: Use the install script (same as CI)
bash test/e2e-v2/script/prepare/setup-e2e-shell/install.sh swctl
export PATH=/tmp/skywalking-infra-e2e/bin:$PATH

# Option 2: Build from source
go install github.com/apache/skywalking-cli/cmd/swctl@<SW_CTL_COMMIT>

yq — YAML processor, used in verify cases:

# Option 1: Use the install script
bash test/e2e-v2/script/prepare/setup-e2e-shell/install.sh yq
export PATH=/tmp/skywalking-infra-e2e/bin:$PATH

# Option 2: brew install yq (macOS)

Other tools (only needed for specific test cases):

All install scripts are at test/e2e-v2/script/prepare/setup-e2e-shell/.

Steps

1. Determine the test case

Resolve the user's argument to a full path under test/e2e-v2/cases/. If ambiguous, list matching directories and ask.

ls test/e2e-v2/cases/<argument>/e2e.yaml

2. Check if rebuild is needed

Compare source file timestamps against the last build:

# OAP server changes since last build
find oap-server apm-protocol -type f \( \
  -name "*.java" -o -name "*.yaml" -o -name "*.yml" -o \
  -name "*.json" -o -name "*.xml" -o -name "*.properties" -o \
  -name "*.proto" \
\) -newer dist/apache-skywalking-apm-bin.tar.gz 2>/dev/null | head -5

# Test service changes since last build
find test/e2e-v2/java-test-service -type f \( \
  -name "*.java" -o -name "*.xml" -o -name "*.yaml" -o -name "*.yml" \
\) -newer test/e2e-v2/java-test-service/e2e-service-provider/target/*.jar 2>/dev/null | head -5

If files are found, warn the user and suggest rebuilding before running.

3. Rebuild if needed (only with user confirmation)

# Rebuild OAP
./mvnw clean flatten:flatten package -Pall -Dmaven.test.skip && make docker

# Rebuild test services
./mvnw -f test/e2e-v2/java-test-service/pom.xml clean flatten:flatten package

4. Run the E2E test

Set required environment variables and run:

export SW_AGENT_JDK_VERSION=8
e2e run -c test/e2e-v2/cases/<case-path>/e2e.yaml

5. If the test fails

Do NOT run cleanup immediately. Instead:

1. Check container logs:

   docker compose -f test/e2e-v2/cases/<case-path>/docker-compose.yml logs oap
   docker compose -f test/e2e-v2/cases/<case-path>/docker-compose.yml logs provider
   

2. Run verify separately (can retry after investigation):

   e2e verify -c test/e2e-v2/cases/<case-path>/e2e.yaml
   

3. Only cleanup when done debugging:

   e2e cleanup -c test/e2e-v2/cases/<case-path>/e2e.yaml
   

6. Manually fire each verify query (fast triage)

The e2e verify retry loop runs in sequence and stops at the first failing case, so a single bad query hides every case after it. When a verify fails, run each verify case directly against the still-running OAP before editing anything — you'll see the real error (bad flag, missing data, wrong expected), not the progress spinner. This is also the right way to author new verify cases: craft the query against live OAP, confirm the actual YAML, then write the expected file.

# Find the host-side port that infra-e2e bound to OAP's container port 12800.
# (Each run picks a new random port; the trigger log prints it too.)
docker ps --filter "name=skywalking_e2e-oap" --format "{{.Ports}}" \
  | grep -oE "[0-9]+->12800" | head -1
# => e.g. 56381->12800

URL=http://localhost:56381/graphql
SWCTL=/tmp/skywalking-infra-e2e/bin/swctl

# Copy the query from e2e.yaml verbatim, then substitute ${oap_host} → localhost
# and ${oap_12800} → the port you just found:
$SWCTL --display yaml --base-url=$URL service ly IOS
$SWCTL --display yaml --base-url=$URL logs list --service-name=MyiOSApp
$SWCTL --display yaml --base-url=$URL metrics exec --expression=service_cpm --service-name=MyiOSApp

When a swctl subcommand rejects a flag (Incorrect Usage: flag provided but not defined: -layer), the e2e config is using syntax the pinned swctl commit doesn't support. Find the right syntax with swctl <cmd> --help and update the e2e config. Common cases encountered:

For queries that don't use swctl (raw curl against /loki/..., Zipkin, PromQL), hit the matching exposed port:

curl "http://localhost:$(docker ps --filter name=skywalking_e2e-oap --format '{{.Ports}}' | grep -oE '[0-9]+->3100' | head -1 | cut -d'-' -f1)/loki/api/v1/labels"

7. UI template changes require a fresh DB

UITemplateInitializer.initTemplate() (in oap-server/server-core) calls uiTemplateManagementService.addIfNotExist(setting) — keyed by the id field in each ui-initialized-templates/**/*.json. Same ID → skipped. So edits to an existing template JSON (adding widgets, relabeling, changing expressions) will not be applied on an already-initialized OAP, even after a container restart, because the old copy still lives in storage.

To pick up dashboard JSON changes:

# Remove both containers — BanyanDB stores state inside the container FS in the
# e2e compose (no named volume), so removing the container wipes state cleanly.
docker rm -f skywalking_e2e-oap-1 skywalking_e2e-banyandb-1

# For compose setups that use a named volume, also:
# docker volume rm <volume-name>

# Then re-run — OAP sees empty storage, loads the new template JSON.
e2e run -c test/e2e-v2/cases/<case>/e2e.yaml

Symptom to watch for: you edit the JSON, rebuild, redeploy — dashboard in the UI still shows the pre-edit layout. That's not a caching bug; that's addIfNotExist doing exactly what its name says.

8. Author the expected YAML from live output

For a new verify case, the workflow is:

1. Fire the query manually (see step 6) and capture the YAML.
2. Pick which fields are meaningful domain values (must match exactly) vs dynamic runtime values (notEmpty / gt / ge). See test/e2e-v2/CLAUDE.md for the decision guide.
3. Write the expected file. If the response is a list, wrap the items in {{- contains . }} ... {{- end }} so ordering and extra actual items don't fail the match.
4. Re-run e2e verify alone (the containers are still up from the previous run); iterate on the expected file without rebuilding.

9. Expected-file authoring traps

These burn CI cycles and pass locally. Each was learned the hard way.

Unquoted content: {{ notEmpty .content }} with : inside the value. Sim-generated or real log content routinely includes colons (POST https://api.example.com/cart failed: 500, HTTP/1.1 500: Internal Server Error). Without quoting, the template renders to invalid YAML (snakeyaml parses failed: as a nested key) and the whole log entry marshals to nil. Symptom: diff shows - nil at every position in the expected logs list vs real maps in actual. Fix: wrap in single quotes — content: '{{ notEmpty .content }}'. Single-quoted YAML preserves : in the scalar; only fails on embedded '. Double quotes also work unless the content has ".

Nested contains with multiple per-element pattern assertions against a varied stream. The template renders the block body once per actual element; when the outer block body has multiple inner contains patterns asserting specific tag key/value pairs, and only some actuals satisfy all the inner patterns, go-cmp with contains can end up comparing [rendered_for_A0, nil, nil, ...] vs [A0, A1, A2, ...] and fail despite contains being permissive on extras. Specifically: outer contains .logs with a single log pattern + inner contains .tags asserting two distinct key/value pairs. On a simulator emitting heterogeneous errors (js + promise + ajax + pageNotFound), only a subset satisfy the inner assertion. Passes locally with 1–2 logs, fails in CI with 6+.

• Keep the outer contains body lenient: field-shape checks (notEmpty, gt), one discriminator tag that every element in the stream carries.
• Cover per-category assertions via separate labeled-metric verify cases, not inside the nested template.
• Rule of thumb: "at least one log exists with the right layer routing" inside the logs expected; per-category coverage via meter_*_count{label=X} verify cases.

Hand-crafted OTLP curl payloads drift from real SDK output. When the upstream SDK ships a published simulator image (mini-program-monitor's sim-wechat / sim-alipay, browser-client-js sim, etc.), prefer driving the e2e with that image in MODE=timed with a bounded DURATION_MS over hand-rolling the OTLP JSON. Hand-crafted payloads miss real-world shape issues: delta-vs-cumulative temporality, label-cardinality surprises, stacktrace formatting variance, attribute key names that changed between SDK versions. Pin to a released tag (v0.4.0), not :latest or HEAD SHA — reproducibility.

timeUnixNano: "0" in an OTLP metric datapoint. The receiver propagates this into MAL's bucket computation and the metric lands in the 1970 time bucket — swctl metrics exec over the "last N minutes" window won't find it. Either use $(date +%s)000000000 at setup time or omit the field if the receiver accepts "now" as default.

Setup-step curl loop with || sleep pattern. The shell line for ... do curl && break || sleep 5; done exits 0 when every attempt connection-refused because the final sleep 5 returns 0. OAP takes ~50 s to start in CI, so all attempts fail before OAP is ready, and the setup step silently succeeds with zero traffic ingested. Fix: curl -sS -f --retry 30 --retry-delay 5 --retry-connrefused --retry-all-errors --max-time 10 ... + set -e at step top.

swctl flag rejected. If a verify case uses a flag the pinned swctl commit doesn't support (service ls --layer vs service ly), the whole case fails 20× before CI gives up. Fire each verify query by hand once before pushing (step 6 above).

Published image cache miss in CI. docker compose pull sometimes hits rate limits or unreachable registries; the test spins until timeout with "dependency failed to start". Look at the CI log for Error response from daemon: pull access denied or manifest unknown. If you see that, pin a different image tag that's definitely published (check docker manifest inspect <tag> locally), not a floating one.

Common test cases