Add debug for merge base logic

Author: sarahchen6

.gitlab-ci.yml

@@ -7,12 +7,18 @@ include:
   - project: 'DataDog/apm-reliability/apm-sdks-benchmarks'
     file: '.gitlab/ci-java-spring-petclinic.yml'
     ref: 'main'
+    rules:
+      - when: never # disable while testing merge base logic
   - project: 'DataDog/apm-reliability/apm-sdks-benchmarks'
     file: '.gitlab/ci-java-insecure-bank.yml'
     ref: 'main'
+    rules:
+      - when: never # disable while testing merge base logic
   - project: 'DataDog/apm-reliability/apm-sdks-benchmarks'
     file: '.gitlab/ci-java-dacapo.yml'
     ref: 'main'
+    rules:
+      - when: never # disable while testing merge base logic
 
 stages:
   - build

.gitlab/scripts/benchmark-compare.sh

@@ -90,12 +90,46 @@ download_job_artifacts() {
   local job_id="$1"
   local output_dir="$2"
   local archive_path="${output_dir}/artifacts.zip"
+  local artifacts_url="${JOBS_API_URL}/jobs/${job_id}/artifacts"
+  local http_code
 
   mkdir -p "${output_dir}"
-  curl --silent --show-error --fail --location \
-    --header "JOB-TOKEN: ${CI_JOB_TOKEN}" \
-    "${JOBS_API_URL}/jobs/${job_id}/artifacts" \
-    --output "${archive_path}"
+
+  http_code="$(
+    curl --silent --show-error --location \
+      --header "JOB-TOKEN: ${CI_JOB_TOKEN}" \
+      --output "${archive_path}" \
+      --write-out "%{http_code}" \
+      "${artifacts_url}"
+  )"
+
+  if [[ "${http_code}" != "200" ]]; then
+    rm -f "${archive_path}"
+    case "${http_code}" in
+      400)
+        echo "ERROR: Bad request for job ${job_id} artifacts (HTTP 400)." >&2
+        ;;
+      401)
+        echo "ERROR: Authentication required for job ${job_id} artifacts (HTTP 401)." >&2
+        ;;
+      403)
+        echo "ERROR: Access forbidden for job ${job_id} artifacts (HTTP 403)." >&2
+        ;;
+      404)
+        echo "ERROR: Artifacts not found for job ${job_id} (HTTP 404). Job may still be running or produced no artifacts." >&2
+        ;;
+      5*)
+        echo "ERROR: Server/network issue while downloading artifacts for job ${job_id} (HTTP ${http_code})." >&2
+        ;;
+      *)
+        echo "ERROR: Failed to download artifacts for job ${job_id} (HTTP ${http_code})." >&2
+        ;;
+    esac
+    if [[ -n "${CI_PROJECT_URL:-}" ]]; then
+      echo "See job details: ${CI_PROJECT_URL}/-/jobs/${job_id}" >&2
+    fi
+    return 1
+  fi
 
   rm -rf "${output_dir}/unzipped"
   mkdir -p "${output_dir}/unzipped"

.gitlab/scripts/get-baseline-commit-info.sh

@@ -18,42 +18,46 @@ fi
 
 readonly PROJECT_API_URL="${CI_API_V4_URL}/projects/${CI_PROJECT_ID}"
 
+log_debug() {
+  echo "[baseline-debug] $*" >&2
+}
+
+response_snippet() {
+  local response="$1"
+  echo "${response}" | tr '\n' ' ' | sed 's/[[:space:]]\+/ /g' | cut -c1-220
+}
+
 get_pipeline_id_for_commit() {
   local commit_sha="$1"
-
-  # Query GitLab API to get pipeline ID that matches the commit SHA
   local api_url="${PROJECT_API_URL}/repository/commits/${commit_sha}"
   local response pipeline_id
+
+  log_debug "query commit endpoint: ${api_url}"
   response="$(curl --request GET --silent --show-error --header "JOB-TOKEN: ${CI_JOB_TOKEN}" "${api_url}" || true)"
   pipeline_id="$(echo "${response}" | grep -o '"last_pipeline"[^}]*"id":[0-9]*' | grep -o '[0-9]*$' | head -1 || true)"
   if [[ -n "${pipeline_id}" && "${pipeline_id}" != "null" ]]; then
+    log_debug "found pipeline_id=${pipeline_id} for commit_sha=${commit_sha}"
     echo "${pipeline_id}"
     return 0
   fi
 
-  echo ""
+  log_debug "no last_pipeline.id for commit_sha=${commit_sha}; response='$(response_snippet "${response}")'"
   return 1
 }
 
-get_branch_head_sha() {
-  local branch="$1"
-
-  # Query GitLab API to get branch head SHA
-  local api_url="${PROJECT_API_URL}/repository/branches/${branch}"
-  local response branch_sha
-  response="$(curl --request GET --silent --show-error --header "JOB-TOKEN: ${CI_JOB_TOKEN}" "${api_url}" || true)"
-  branch_sha="$(echo "${response}" | grep -o '"commit"[^}]*"id":"[a-f0-9]\{40\}"' | sed -E 's/.*"id":"([a-f0-9]{40})".*/\1/' | head -1 || true)"
-
-  echo "${branch_sha}"
-}
-
 get_latest_pipeline_id_for_branch() {
   local branch="$1"
-  local api_url="${PROJECT_API_URL}/pipelines?ref=${branch}&status=success&order_by=updated_at&sort=desc&per_page=1"
+  local api_url="${PROJECT_API_URL}/pipelines?ref=${branch}&order_by=id&sort=desc&per_page=1"
   local response pipeline_id
 
+  log_debug "query pipelines endpoint: ${api_url}"
   response="$(curl --request GET --silent --show-error --header "JOB-TOKEN: ${CI_JOB_TOKEN}" "${api_url}" || true)"
   pipeline_id="$(echo "${response}" | grep -o '"id":[0-9]*' | head -1 | grep -o '[0-9]*' || true)"
+  if [[ -n "${pipeline_id}" ]]; then
+    log_debug "found latest pipeline_id=${pipeline_id} for branch=${branch}"
+  else
+    log_debug "no pipeline found for branch=${branch}; response='$(response_snippet "${response}")'"
+  fi
   echo "${pipeline_id}"
 }
 
@@ -73,21 +77,24 @@ BASELINE_PIPELINE_ID=""
 BASELINE_SOURCE="merge_base"
 FALLBACK_TO_MASTER="false"
 
+log_debug "resolved merge_base_sha='${MERGE_BASE_SHA:-<empty>}' target_branch='${TARGET_BRANCH}'"
+
 if [[ -n "${MERGE_BASE_SHA}" ]]; then
   BASELINE_PIPELINE_ID="$(get_pipeline_id_for_commit "${MERGE_BASE_SHA}" || true)"
 fi
 
 if [[ -z "${BASELINE_PIPELINE_ID}" ]]; then
   FALLBACK_TO_MASTER="true"
   BASELINE_SOURCE="${TARGET_BRANCH}"
+  log_debug "merge-base pipeline not found; falling back to branch='${TARGET_BRANCH}'"
 
   BASELINE_SHA="$(git rev-parse "origin/${TARGET_BRANCH}" 2>/dev/null || true)"
-  if [[ -z "${BASELINE_SHA}" ]]; then
-    BASELINE_SHA="$(get_branch_head_sha "${TARGET_BRANCH}" || true)"
+  log_debug "resolved branch sha from local git: '${BASELINE_SHA:-<empty>}'"
+  if [[ -n "${BASELINE_SHA}" ]]; then
+    BASELINE_PIPELINE_ID="$(get_pipeline_id_for_commit "${BASELINE_SHA}" || true)"
   fi
-  BASELINE_PIPELINE_ID="$(get_pipeline_id_for_commit "${BASELINE_SHA}" || true)"
 
-  # use latest successful branch pipeline
+  # Final fallback: use latest branch pipeline
   if [[ -z "${BASELINE_PIPELINE_ID}" ]]; then
     BASELINE_PIPELINE_ID="$(get_latest_pipeline_id_for_branch "${TARGET_BRANCH}" || true)"
   fi