Investigate Salesforce failures

Own the loop from a symptom to evidence, the smallest supported correction and verification within the requested scope. Start with bulk sync and local content search.

Establish the corpus

Infer workspace, intended org and clues from the request/project. Ask only for missing context that matters. Record IDs, order numbers, exception fragments and class names can locate the transaction without a log ID.

For current org activity run sf electivus doctor --target-org my-org --json, then sf electivus log sync --target-org my-org --json in the workspace. Inspect result.status, failed, failures and orgLogsRoot: outer status: 0 does not prove sync completeness. Subsequent syncs are incremental. --force-full reconciles retained history missing behind a checkpoint; do not use it for every search. Diagnose persistent sync errors before retrying.

For supplied/offline logs, read directly without requiring auth or installing the CLI. Older versions expose apexlogsRoot and safeTargetOrg through log status; use public --help to confirm unfamiliar capabilities. Software installation or updates must be in scope.

Find the execution

Read references/investigation.md for recipes and interpretation. Find candidates first with local tools, then inspect relevant context:

rg --no-ignore -l -F -g '*.log' -- 'business-id' '/absolute/orgLogsRoot'