Signals · 3 min
What SDK-native APM actually is
Application performance from inside your process — transactions, RED, and release health — without a host agent or an observability maze.
Learn
Learn the product, not the jargon.
SDK-native application performance — transactions, RED, traces, Web Vitals, and release health — explained against the dashboard you will actually use. Not host metrics, profiling, or a log platform.
Signals
What SDK-native APM measures — traces, waterfalls, RED, and sampling — and why that is not host monitoring.
Signals · 3 min
Transactions, traces, and spans
A trace is one request’s path. A span is a timed unit of work. A transaction is the root span Talaria meters.
Signals · 3 min
How to read a request waterfall
A waterfall is the trace drawn as time. The critical path is the longest chain of parent-to-child work — that is usually the fix.
Signals · 3 min
RED metrics: rate, errors, duration
Three numbers that describe a service from the outside: how often it is called, how often it fails, and how long it takes.
Signals · 3 min
Sampling you can explain
Keep every failing transaction. Sample a tenth of the rest. Bill the root, not every query. That is the whole policy.
In the request
The work inside a transaction: databases, outbound HTTP, Web Vitals, and W3C context across services.
In the request · 3 min
Slow queries and N+1
Most “the database is slow” incidents are the application asking the same question too many times — or once, too late.
In the request · 3 min
External dependencies
Every outbound call is a peer: HTTP hosts, databases, Redis, queues. Watch their error rate and p95 the same way you watch your own routes.
In the request · 3 min
Web Vitals from real users
LCP, INP, and CLS are how Chrome defines a usable page. Talaria records them from the browser SDK as spans on real sessions — not a lab crawl.
In the request · 3 min
Distributed tracing with W3C Trace Context
One trace ID from the browser, through PHP, into the query that failed — carried on a standard header named traceparent.
Triage loop
How performance joins issues, replays, releases, alerts, and AI so a slow request becomes something you can own.
Triage loop · 3 min
Release health
A version is healthy when sessions survive it and new issues do not explode. Tag events with a release, then read crash-free rate instead of guessing from git tags.
Triage loop · 3 min
From error to trace to replay
Performance without grouping is a chart. Grouping without traces is a stack. Replay without either is a video. Talaria joins all three on one issue.
Triage loop · 3 min
Performance alerts and AI on traces
Alerts notice when RED crosses a line you set. AI reads the traces and issues you already have. Neither is an autonomous SRE.
Ready to instrument?
Tracing is off until you opt in. Install an official SDK, set enableTracing, and open Performance — child spans are included, not billed.
Prefer a feature inventory? See Features.
See it on a real project
Start free on Developer — 10,000 included transactions a month, with events and replays metered separately.