APM agent
Automatic instrumentation for Java, .NET, Node.js and Python with one command, on Linux and Windows.
Argunis/Operations/APM
Follow every request inside.
With the APM agent, an OpenTelemetry SDK or eBPF, Argunis sees every request in your applications: which service handled it, which database it queried and what made it slow or fail.
APM · 23 services
Requests / min18.420
Errors0,8%
p95 latency284 ms
Services
The waterfall of every request
Filter traces by service, endpoint, status, duration or attribute, and read the waterfall to see exactly which step was slow or failed, down to the SQL query.
- Filters by service, endpoint and attribute
- Waterfall with every span
- Logs from the same trace
- Browser linked to the back end
Trace · POST /payments · 1.9 sslow query
Who calls whom
The dependency map shows every service with its health and the traffic of each connection, so you find where a problem starts. Live, the last 15 minutes, also on the NOC screen.
- Map from real traces
- Health, traffic, latency and errors per connection
- Live APM full screen
- Services linked to the inventory
Dependencies · last 15 min
- 1web-store
- 2api-payments
- 3auth
- 4PostgreSQL
- 5Kafka
api-payments → PostgreSQL accounts for 74% of the checkout flow latency.
Grouped errors and alerts
Exceptions are grouped by cause. Mark them as new, known, resolved or ignored, and “Declare fix” closes their incident. Latency and error alerts open incidents in the Core.
- Exceptions grouped by cause
- States: new, known, resolved, ignored
- Adjustable thresholds per service
- Retention and removal of personal data
Errors · api-payments · 24 h
TimeoutExceptionPaymentRepository.java:11894
NullPointerExceptionCoupon.java:427
Everything APM includes
SDK with token
Your application sends straight to Argunis: also PHP, Go and Ruby, containers and clouds.
eBPF
Observes an application from the operating system without touching it: HTTP, gRPC, databases and queues.
Browser
Page loads, fetch and XHR from the front end, joined to the back-end trace.
Services
Grouped by system, with health, traffic, errors, latency and inventory link.
Traces
Search by service, endpoint, status, duration or attribute, with its waterfall.
Dependency map
Who calls whom, with the traffic and health of each connection.
Errors
Grouped exceptions, with states and incident closing.
Retention and privacy
Days of traces, logs and metrics, and personal data removed before storage.
ApplicationsFrequently asked questions
Which languages are supported?
With the APM agent: Java, .NET Framework, .NET, Node.js and Python. With the SDK also PHP, Go, Ruby and the browser. With eBPF any Linux application can be observed without changes.
Do I have to change my code?
Usually not. The agent instruments the application at start-up and eBPF observes it from the operating system. Go needs a few lines at start-up.
What about personal data in traces?
You define what is removed before storage and how many days traces, logs and metrics are kept.
See the failure before your customers do.
We show you Argunis on your own operations and answer your questions in one call.