The best way to control Claude Code, Codex and more from any computer or phone

main
mdx 98 lines 3.13 KB
Raw
1 ---
2 title: Axiom
3 description: Observability as resources — OTEL datasets, ingest tokens, monitors, notifiers, and dashboards declared next to the code that emits the data.
4 ---
5
6 Axiom is an event and observability store for logs, traces, and
7 metrics. Effect already emits OpenTelemetry, and with alchemy the
8 receiving end is code too: datasets, ingest tokens, monitors,
9 notifiers, and dashboards live in the same Stack as the Workers and
10 Functions they observe, so a threshold change is a reviewable diff.
11
12 New here? [Set up credentials](/axiom/setup) first.
13
14 ## Resources
15
16 A `Dataset` is the top-level container. Pick a `kind` per OTEL
17 signal — it exposes the OTLP endpoints as outputs:
18
19 ```typescript
20 const traces = yield* Axiom.Dataset("traces", { name: "app-traces", kind: "otel:traces:v1" });
21 const logs = yield* Axiom.Dataset("logs", { name: "app-logs", kind: "otel:logs:v1" });
22 const metrics = yield* Axiom.Dataset("metrics", { name: "app-metrics", kind: "otel:metrics:v1" });
23 ```
24
25 :::caution
26 Renaming a dataset or changing its `kind` is a **replacement** — the
27 old dataset and all its events are deleted. See
28 [Datasets & Ingest](/axiom/data/ingest) for details.
29 :::
30
31 An `ApiToken` mints a least-privilege ingest credential for your
32 runtime to ship with:
33
34 ```typescript
35 const ingest = yield* Axiom.ApiToken("ingest", {
36 name: "prod-ingest",
37 datasetCapabilities: {
38 "app-traces": { ingest: ["create"] },
39 },
40 });
41 ```
42
43 `Monitor` + `Notifier` turn queries into alerts — see
44 [Monitors & Alerting](/axiom/guides/alerting) for threshold and
45 anomaly monitors:
46
47 ```typescript
48 const slack = yield* Axiom.Notifier("ops-slack", {
49 name: "ops-channel",
50 properties: {
51 slack: { slackUrl: process.env.SLACK_WEBHOOK_URL! },
52 },
53 });
54
55 yield* Axiom.Monitor("panics", {
56 name: "Service panic",
57 type: "MatchEvent",
58 aplQuery: `['app-logs'] | where message contains "panic:"`,
59 intervalMinutes: 1,
60 rangeMinutes: 1,
61 notifierIds: [slack.id],
62 });
63 ```
64
65 Beyond alerts, `Dashboard` declares charts as code, `View` saves an
66 APL query as a virtual dataset, and `Annotation` marks deploys on
67 your charts — see [Dashboards as Code](/axiom/guides/dashboards).
68 `VirtualField` derives computed fields on a dataset at query time.
69
70 ## Compose with your cloud
71
72 Wire the dataset's OTLP endpoints and the ingest token into your
73 Worker or Lambda's environment, and Effect's OpenTelemetry output
74 flows into Axiom with no code changes — see
75 [Observability](/testing/observability) for the exporter-as-a-Layer
76 pattern:
77
78 ```typescript
79 vars: {
80 OTEL_EXPORTER_OTLP_TRACES_ENDPOINT: traces.otelTracesEndpoint,
81 OTEL_EXPORTER_OTLP_LOGS_ENDPOINT: logs.otelLogsEndpoint,
82 },
83 ```
84
85 For a full end-to-end walkthrough — dataset, token, Worker wiring,
86 and a monitor on top — follow
87 [Observability with Axiom](/cloudflare/observability/axiom-observability).
88
89 ## Reference
90
91 [Dataset](/providers/axiom/dataset) ·
92 [ApiToken](/providers/axiom/apitoken) ·
93 [Monitor](/providers/axiom/monitor) ·
94 [Notifier](/providers/axiom/notifier) ·
95 [Dashboard](/providers/axiom/dashboard) ·
96 [Annotation](/providers/axiom/annotation) ·
97 [View](/providers/axiom/view) ·
98 [VirtualField](/providers/axiom/virtualfield)