real projects · hybrid · advanced
Observability Scripting
Build small probes, structured metrics exporters, and alert budgets that humans respect.
5 weeks · 30 guided hours · rolling · 17,200 THB (informational)
Tool stack
PythonOpenTelemetryPrometheus
Description
You will emit OpenTelemetry-friendly spans from Python services, craft SLI queries, and tune alert noise using Thai holiday calendars as a real constraint. Labs include translating Thai error strings without breaking UTF-8 pipelines.
What is included
- Redaction helpers for log lines with citizen IDs
- Metric naming conventions that survive reorgs
- SLO drafts with error budgets tied to business hours
- Synthetic checks with empathetic failure copy
- Notebook dashboards for ad-hoc investigations
- Sampling strategies for cost control
Outcomes
- Ship a probe suite with documented maintenance owner
- Draft an SLO doc reviewed in mentor session
- Demonstrate alert suppression during planned maintenance windows
FAQ
Vendor-specific agents?
We discuss patterns; vendor installers are your IT team's scope.
Math prerequisites?
Comfort with percentiles and basic algebra is enough.
Not covered?
Full APM negotiation with procurement teams.
Experience notes
“Redaction helper snippet went straight into our staging logger—mentor flagged one edge case on merged log lines.”
“Observability Scripting labs gave us alert copy that front desk staff understood during the dry run.”