Skip to content

chore(deps): update the pinned @evlog/cli to v0.12.0 - #10

Open
renovate[bot] wants to merge 1 commit into
mainfrom
renovate/evlog-cli-0.x
Open

renovate[bot] wants to merge 1 commit into
mainfrom
renovate/evlog-cli-0.x

Conversation

@renovate

@renovate renovate Bot commented Oct 12, 2026 •

Copy link
Copy Markdown

This PR contains the following updates:

Package Change Age Confidence
@evlog/cli (source) 0.8.0 → 0.12.0 age confidence

Release Notes

evloghq/evlog (@​evlog/cli)

v0.12.0

Compare Source

Minor Changes
  • #​1075 8ee1a17 Thanks @​HugoRCD! - evlog map now records what each entry point's wide event carries. Every route in evlog.map.json and in --json gains an event key with the dotted paths log.set() writes (user.id, cart.items) and the literal action of each log.audit() call. Comparing it between two scans shows which fields a change removes from your production logs. Routes that set nothing have no event key, and the existing keys are unchanged.
Patch Changes
  • #​1083 9621ea4 Thanks @​HugoRCD! - evlog agents and evlog init no longer delete or duplicate text in an AGENTS.md whose evlog markers are not one clean pair. A file that lost <!-- evlog:end --> got a second block on one run and lost the section below it on the next. Both commands now stop with cli.AGENTS_MARKERS_UNPAIRED, exit 1 and leave the file as it is. A marker quoted in a sentence or a code fence is read as text.

    evlog init also changes in four places:

    • A Nitro config that imports something else from evlog/nitro or evlog/nitro/v3 gets import evlog along with modules: [evlog()]. It got the module entry with no import.
    • The seeded errors.ts and audit.ts parse when a service or a route starts with a digit (_3dShopErrors, _2FA_PASSWORD_CREATED) or a route holds a wildcard.
    • A failed npx skills add exits 1, and the AGENTS.md block no longer points at the skill that was not installed. evlog agents already did both.
    • A file that cannot be written stops the run with cli.WRITE_FAILED, which names the file and the ones written before it, under --json too. evlog agents does the same. Both printed a stack trace and left stdout empty.
  • #​1128 e1e33d0 Thanks @​HugoRCD! - Two fixes to evlog init and evlog agents.

    • On Windows, evlog init --apps apps/web sets up the app. Workspace app paths were built with backslashes there, so the forward-slash form was rejected as unknown, and the JSON app values and the per-app headings carried backslashes. They use forward slashes on every platform, and --apps apps\web still selects the app.
    • When npx skills add fails, the reported error is the last line it wrote before npm's update notice. Before, a failed first run could report npm notice as the error.
  • #​1101 84873d5 Thanks @​HugoRCD! - The CLI says what is wrong with the environment it is started in, and its output behaves in a pipe.

    • On Node older than 22, every command stops with @evlog/cli needs Node.js 22.0.0 or later, and this is Node.js 20.20.2 and exit code 1. Before, evlog map failed there with a syntax error about globSync, while doctor, logs and --version still ran on a Node the package does not support. evlog doctor no longer has a failing node check or a cli.NODE_TOO_OLD finding, since it never starts there.
    • Next to an evlog older than the peer range, or with none installed, every command stops with the version it needs, the version and path it found, and exit code 1. Before, Node reported a missing export as a syntax error.
    • A package.json that does not parse in the working directory no longer stops the launcher with ERR_INVALID_PACKAGE_CONFIG.
    • --help prints to stdout, so evlog --help | grep map works. Usage shown with a usage error stays on stderr. Flags are listed the way they are typed (--min-score, -f, --follow), and positional arguments have their own section.
    • Colour follows the stream the report is written to: evlog doctor 2> report.txt writes plain text from a terminal. FORCE_COLOR=0 and TERM=dumb turn colour off, and the header rule uses 24-bit colour only where the terminal supports it.
    • A reader that closes the pipe early, as in evlog logs --json | head, ends the run quietly instead of with an unhandled EPIPE error.
  • #​1059 4466f57 Thanks @​HugoRCD! - Every command now stops on a flag it does not declare, and on a value flag passed with no value, with exit code 2. evlog map --min-scor 90 used to run with no gate and exit 0, and evlog init --dryrun wrote files. A flag written before the command name is checked too: evlog --dry-run init used to run init for real. A script that carries a mistyped flag today starts failing: fix the spelling.

    The other usage errors exit 2 as well, as the README documents: an unknown command, an invalid --min-score, --framework, --format or --limit, and evlog map <entry> for an entry point that does not exist. An empty --min-score, which is what an unset shell variable expands to, is one of them. A bare --baseline no longer takes the flag after it as its value: evlog map --baseline --min-score 80 and evlog map --baseline --cwd apps/web read both flags. A map error is printed as JSON under --format json, as it already was under --json.

  • #​1085 03b63a8 Thanks @​HugoRCD! - evlog logs reads a log of any size without its memory growing with it. Every view held each matching event until the end of the run: an 800 MB log took 1.2 GB with --limit 1. A view now keeps only the --limit events it shows, and stats keeps one duration per event for the percentiles. The output of every view is unchanged.

  • #​1083 9621ea4 Thanks @​HugoRCD! - evlog logs applies every --where when the flag is repeated. Only the last clause applied, so --where path=/a --where n=5 returned events on any path. An empty --where, which is what an unset shell variable expands to, is now a usage error with exit 2.

    An event whose duration, level or service is not text no longer crashes every view, --json included.

    A --where value is a number only when it spells one exactly. zip=123 no longer matches "00123", a 19-digit id no longer matches its neighbours, and amount>5000 no longer matches a field that holds text such as "n/a". Write the value as the field holds it: zip=00123.

  • #​1113 6011cd1 Thanks @​HugoRCD! - evlog map scores each entry point on its own handler instead of on the whole file.

    • In a file that declares several entry points (Express, Hono and Fastify routes, the methods of a Next.js route.ts, server actions, TanStack Start methods), one instrumented handler used to pass every other one in the file, and each route listed the same event.fields. Each entry point is now read from its handler and the functions, classes and constants of the same file it names. An import such as stripe marks only the handlers that use it as sensitive.
    • request.log.info() in a Fastify app that never registers evlog's plugin no longer passes wide-event: that is Fastify's own logger. req.log and request.log count when the app registers evlog's middleware, or when the call is set, audit or fork. request.evlog is recognised.
    • c.get('log').set() called without binding the logger first is credited in Hono.
    • evlog map --all names each row of a shared file by its route (index.ts · GET /health). evlog map <file> lists the other entry points of the file, and evlog map "POST /orders" inspects one by method and path.

    Scores drop in files where an instrumented route was covering for the others, so the rule-set version moves to 3. --baseline against a map written by an earlier CLI exits 2 with the version mismatch: run evlog map once and commit the new evlog.map.json.

  • #​1059 4466f57 Thanks @​HugoRCD! - evlog map no longer scores a tree it only partly read.

    • A file with a syntax error, or one the scan cannot read, used to lose its routes, which raised the score. The report now warns <file> failed to parse and counts the file as one failing entry point. With --min-score, map.minScore or --baseline, the run stops with cli.MAP_PARSE_FAILED, exits 1, and leaves evlog.map.json as it was. List the file under map.ignore when it is not app code.
    • Directories named build, dist and coverage are skipped at the project root only, so server/api/build/index.get.ts is scanned.
    • Two entry points with the same method and path, such as get('/') in two routers, get distinct ids. --baseline reported regressions on an unchanged tree when they shared one. The id of every other entry point is unchanged, so a committed map stays comparable, and an entry point is still compared with itself when a second one with its method and path appears or goes. If it also moved to another file in that change, it is compared with the entry point that has its method and path and lost the most checks, then the lowest-scoring one.
    • .js, .mjs and .cjs files are parsed with JSX, so a Next.js page.js is checked instead of reported as n/a.

    The first and the last of these can change a verdict on code you did not touch, so the rule-set version moves to 2. --baseline against a map written by an earlier CLI exits 2 with the version mismatch: run evlog map once and commit the new evlog.map.json.

  • #​1059 4466f57 Thanks @​HugoRCD! - New telemetry.fail(code?) for a command that handles its own failure: it prints the error, sets an exit code and returns, so the run used to be recorded as outcome: "success". Call it where the error is handled and the event carries outcome: "error" with code as its errorCode.

    The evlog CLI calls it for every error it reports, so evlog logs with no log directory, evlog config on an invalid file and a mistyped flag are sent as errors with their cli.* code. A failed --min-score or --baseline gate stays a successful run with mapGateFailed: true.

  • #​1059 4466f57 Thanks @​HugoRCD! - Telemetry now does what its docs say.

    • The run that opts out is no longer queued: telemetry disable, or disableTelemetry() called inside a run, used to purge the outbox and then write its own event to it. disableTelemetry() turns off every handle created for the tool, and a flush() already sending stops at the batch in flight.
    • EVLOG_TELEMETRY_DEBUG=1 prints each run's event as the run ends and makes no request. The event is not queued either. It used to print at the next flush, and that flush sent the events.
    • An endpoint that never answers no longer keeps the process alive for ten seconds. The drain closes its connection at the 400 ms timeout. It no longer follows redirects, so endpoint has to be the final URL.
    • DO_NOT_TRACK=true, yes and on opt out like 1, and EVLOG_TELEMETRY=false, no and off like 0.
    • withTelemetry() records only the flags a command declares. An undeclared flag used to be sent by name, and a positional as <set>.
    • New bin option: the command people type, when it is not name. The first-run notice prints <bin> telemetry disable, so the evlog CLI's notice now says evlog telemetry disable, not evlog-cli telemetry disable.
    • evlog telemetry status lists the allowlisted init*, map* and logs* values. The fields evlog logs records are documented.
  • Updated dependencies [777e3bb, 4466f57, 4466f57]:

v0.11.2

Compare Source

Patch Changes
  • #​990 b4c170f Thanks @​HugoRCD! - Escape control characters in evlog logs output. Before, a logged value such as a request path containing \r\x1b[2K was printed to the terminal as is, so it could erase the line it was on and print a fake log line in its place. The one-line view, the full view of one event, its FIELDS block and the stats table now print control characters as \r, \n and \u001b, the way evlog's pretty output does since its last patch. Tabs are kept. --json output is unchanged.

  • #​992 6f6e9db Thanks @​HugoRCD! - evlog map now reads a Vue route file whose <script> block closes with </script > or another closing tag that has whitespace before >. Before, it found no script in that file and skipped it.

v0.11.1

Compare Source

Patch Changes
  • #​945 17b0411 Thanks @​evlogai! - evlog init now writes evlog.config.ts with process.env.NODE_ENV in place of import.meta.dev for Nuxt, Nitro and TanStack Start, the same condition Next, Hono, Express and Fastify already used. The generated file no longer reports Property 'dev' does not exist on type 'ImportMeta' in a fresh Nuxt project before nuxi prepare has run. The production drain is still selected in a Nuxt production build, because Nitro inlines NODE_ENV there. Existing evlog.config.ts files are not rewritten; change the condition by hand to pick this up.

  • #​949 20aaa2c Thanks @​evlogai! - evlog logs and the NuxtHub log viewer now show the HTTP status of an error from error.status, falling back to error.statusCode for files written before v3. Before, a v3 error event rendered no status, and a 2.x error created with only status (for example a plain h3 error) rendered none either.

v0.11.0

Compare Source

Minor Changes
  • #​882 cf3a43b Thanks @​evlogai! - evlog init offers Grafana Loki and ClickHouse as production destinations. Choosing one wires createLokiDrain() from evlog/loki or createClickHouseDrain() from evlog/clickhouse, and asks for LOKI_ENDPOINT or CLICKHOUSE_ENDPOINT.

  • #​881 68d709a Thanks @​evlogai! - evlog init offers Sazabi as a production destination. Choosing it wires createSazabiDrain() from evlog/sazabi and asks for SAZABI_INTAKE_URL, the intake URL of an OpenTelemetry log stream in Sazabi.

Patch Changes
  • #​887 71b2d77 Thanks @​evlogai! - evlog doctor and evlog init now recognise an app that reaches evlog only through a wrapper package, such as a shared @acme/logger that calls createEvlog, the same way evlog map already does. doctor reports evlog as reached through that package (for example "evlog via @acme/logger") instead of "not found in this project", drops the pnpm add evlog hint, and no longer warns that the framework still needs evlog. init stops in that case and says evlog is already wired through the wrapper, rather than offering to install evlog and scaffold a second setup next to it. doctor --json reports the same result.

  • #​869 99189f3 Thanks @​HugoRCD! - evlog map recognises evlog helpers imported from a package that wraps evlog, such as a shared @acme/logger that calls createEvlog in a workspace. A dependency counts when it declares evlog and forwards evlog's exports from one of its exports entries, whether as source or as built output. Handlers that import useLogger or withEvlog from it are no longer reported as dark events, and an app that reaches evlog only through that package counts as having evlog installed. In a "use server" file, an export built by a call, such as export const checkout = withEvlog(async () => …), is now listed as a server action.

  • #​895 6203a07 Thanks @​evlogai! - The published packages no longer ship .map files. They pointed to src/, which is not part of the package, so they only added to the install size. evlog drops from 340 files (about 2.09 MB unpacked) to 183 files (about 905 KB). No runtime or API change.

  • Updated dependencies [6203a07]:

v0.10.0

Compare Source

Minor Changes
  • #​807 7e6282a Thanks @​HugoRCD! - evlog config reads the options passed to the evlog module in nuxt.config.ts or nitro.config.ts, which override evlog.config.ts at runtime. A note under the file name says so, and each setting they replace shows the value that applies with its line, and the value of the file under it. --json adds moduleOptions and an overrides: { value, source } field on those settings.

  • #​806 704c311 Thanks @​HugoRCD! - evlog map and evlog logs read evlog.config.ts (or .mts, .js, .mjs), the nearest one from the package up to the workspace root, without running it. map.rules turns checks off or back on for every entry point, map.ignore leaves entry points out by file glob and --json reports how many as ignored, map.minScore and map.baseline gate like --min-score and --baseline, and logs.dir and logs.limit set the defaults of --dir and --limit. A flag wins over the config. The config can extend a local file or a published preset, one level deep. evlog config prints the settings that apply, grouped into service, sampling, redaction, pipeline and CLI, with the file and line each one comes from and evlog's defaults where the file says nothing, --json for the same as JSON, and evlog doctor reports a config it cannot read. A misspelt setting, an extends written as a path instead of an imported config, a value under map or logs computed at runtime, or turning off wide-event or context stops the command with a cli.CONFIG_* error. The CLI now needs evlog 2.31.0 or later.

  • #​808 09b8d50 Thanks @​HugoRCD! - evlog init writes evlog.config.ts with the service name, drains, enrichers and sampling you pick, dev and production branched in one place. Nuxt and Nitro load it through the module, and on Next.js, Hono, Express and Fastify the generated evlog.ts imports it. In a workspace with a root evlog.config.ts, the app config extends it and leaves out what the root already sets. A config you already have is read and never rewritten: a setting it lacks becomes a manual step, and the service it names becomes the default answer.

  • #​833 62637e9 Thanks @​evlogai! - A global evlog now hands off to the @evlog/cli installed in the project it runs in, so a project always runs the CLI version it pins. In a directory with no install, the global copy runs as before.

  • #​831 b7b8c8b Thanks @​evlogai! - evlog agents and evlog init now install the agent skills globally by default, for universal (the shared .agents/skills directory that Codex and the other agents read) and claude-code (a symlink into it). Pass --local to install into the project instead, with .agents/skills and .claude/skills at the project root.

    Breaking: --global is removed, since global is now the default. evlog agents no longer writes CLAUDE.md. Most agents read AGENTS.md directly, and Claude Code does too unless a CLAUDE.md sits above it. Add a one-line @AGENTS.md yourself if you want one; an existing CLAUDE.md is left alone.

Patch Changes
  • #​824 75a32d8 Thanks @​evlogai! - Print the "evlog skills already installed" line once in evlog init and evlog agents. The plan already lists it, so the status line below now shows only the npx skills update refresh command.

  • #​822 2d7503a Thanks @​evlogai! - Fix two evlog map results. useLogger(event).set({ … }) written inline, without binding the logger to a variable, is now credited as request context, so the context rule no longer reports the handler as missing log.set(). A project with no entry points to score is now graded unscored ("nothing to scan") instead of a free "excellent" 100, and --min-score exits 1 in that case because there is nothing to meet the threshold against.

  • #​823 b603f28 Thanks @​evlogai! - evlog map no longer draws a score for an unscored scan. A project with no entry points to score showed a full banner reading 100 and a full bar, next to "nothing to scan". The banner now draws -- with an empty bar, and --all shows -- in place of 100/100. The JSON output and the --min-score gate were already correct and are unchanged.

  • #​836 f3613ee Thanks @​HugoRCD! - Redact secrets by key name. A new secrets built-in, on by default with the other built-ins, replaces the value of any key ending in password, passwd, secret or token (case-insensitive, at any depth) with replacement. So password, newPassword, client_secret and accessToken no longer reach production logs, while counters such as totalTokens and tokensPerSecond are kept. Values are replaced in full, booleans and null are left alone, and listing builtins without 'secrets' turns it off. evlog config lists secrets among the built-ins.

v0.9.0

Compare Source

Minor Changes
  • #​773 a053adb Thanks @​HugoRCD! - evlog map --json now carries grade next to map and summary, the same word the report prints, so a consumer no longer recomputes it from the score. --baseline-label <text> names the baseline in the report and the --format github annotations when the map was copied to a temp file, so CI can say regressed against main instead of printing a path.

  • #​774 5ca8bd4 Thanks @​HugoRCD! - evlog logs reads the wide events the fs drain wrote to .evlog/logs: the last 50 (evlog logs), the failures (evlog logs errors), the slowest (evlog logs slow --over 1s), one request in full by id (evlog logs <requestId>, a UUID prefix is enough), or the shape of the traffic (evlog logs stats: per route, status class and level). Filters compose with every view: --since 15m, --until, --level error,fatal, --path, --status 5xx, --where payment.amount>5000 on any field of the event (=, !=, >, >=, <, <=, ~regex, present, !absent, repeatable), --limit, --dir. -f follows new events like tail -f; --json returns the events as JSON. It finds the project's log directory the way doctor does, reads every app of a workspace when the root has none, reads the memory drain's dev endpoint with --url (a follow there survives the app restarting, and says so when the endpoint stays quiet), handles both the compact and the pretty layout, and never writes.

Patch Changes

Configuration

📅 Schedule: (in timezone UTC)

  • Branch creation
    • At any time (no schedule defined)
  • Automerge
    • At any time (no schedule defined)

🚦 Automerge: Disabled by config. Please merge this manually once you are satisfied.

♻ Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

🔕 Ignore: Close this PR and you won't be reminded about this update again.


  • If you want to rebase/retry this PR, check this box

This PR was generated by Mend Renovate. View the repository job log.

@coderabbitai

coderabbitai Bot commented Oct 12, 2026 •

Copy link
Copy Markdown

Important

Review skipped

Bot user detected.

To trigger a single review, invoke the @coderabbitai review command.

Run configuration

  • Source mode: Auto
  • Repository: Unavailable
  • Branch: Unavailable
  • File: No configuration file from this repository applied
  • Commit: Unavailable
  • Layers: defaults
  • Review profile: CHILL
  • Plan: Advanced
  • Run ID: d3b35221-3b74-49b2-82fa-0adb82481753

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

  • Trigger review
  • Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

Share

@github-actions

github-actions Bot commented Oct 12, 2026 •

Copy link
Copy Markdown

Observability score · 64

needs work across 3 packages · unchanged against main · no regression

Package Score Δ Instrumented Partial Dark Gate
test/fixtures/express-app (express) 73 needs work 0 1 0 1 passed
test/fixtures/hono-app (hono) 73 needs work 0 1 0 1 passed
test/fixtures/nuxt-app (nuxt) 64 needs work 0 2 0 2 passed

Note

No min-score set. Add min-score: 64 to keep today's score from dropping.

Fix first (4)

evlog map v0.12.0 · how the score works · what each check expects

@renovate
renovate Bot force-pushed the renovate/evlog-cli-0.x branch from 830932e to 26dea5b Compare October 17, 2026 15:48
@renovate renovate Bot changed the title chore(deps): update the pinned @evlog/cli to v0.10.0 chore(deps): update the pinned @evlog/cli to v0.11.0 Oct 17, 2026
@renovate
renovate Bot force-pushed the renovate/evlog-cli-0.x branch from 26dea5b to d10d627 Compare October 30, 2026 08:21
@renovate renovate Bot changed the title chore(deps): update the pinned @evlog/cli to v0.11.0 chore(deps): update the pinned @evlog/cli to v0.11.1 Oct 30, 2026
@renovate renovate Bot changed the title chore(deps): update the pinned @evlog/cli to v0.11.1 chore(deps): update the pinned @evlog/cli to v0.11.2 Nov 7, 2026
@renovate
renovate Bot force-pushed the renovate/evlog-cli-0.x branch from d10d627 to d31484c Compare November 7, 2026 15:16
@renovate
renovate Bot force-pushed the renovate/evlog-cli-0.x branch from d31484c to ea173fe Compare November 20, 2026 18:37
@renovate renovate Bot changed the title chore(deps): update the pinned @evlog/cli to v0.11.2 chore(deps): update the pinned @evlog/cli to v0.12.0 Nov 20, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

0 participants