How to Use Big:eye Pro: Tips, Tricks, and Setup
What Big:eye Pro does
Big:eye Pro is a monitoring and observability tool for applications and infrastructure that collects metrics, traces, and logs to help you detect issues, optimize performance, and troubleshoot faster.
Quick setup (assumes Linux server + web UI)
-
Sign up and create a project
- Visit the Big:eye Pro web console and create an account (or sign in to your organization). Create a new project/environment for the service you’ll monitor.
-
Install the agent
- On the target host, download the official Big:eye Pro agent package (RPM/DEB or tarball).
- Install and start the agent as a service:
bash
# Debian/Ubuntu (example) sudo dpkg -i bigeyepro-agent.deb sudo systemctl enable –now bigeyepro-agent # RHEL/CentOS (example) sudo rpm -ivh bigeyepro-agent-.rpm sudo systemctl enable –now bigeyepro-agent - Confirm the agent is running:
bash
sudo systemctl status bigeyepro-agent
-
Authenticate the agent with an API key
- In the Big:eye Pro console, create an API key for the project.
- Place the key in the agent config (commonly /etc/bigeyepro/agent.conf) or export as an env var:
bash
export BIGEYE_API_KEY=“your_api_keyhere” sudo systemctl restart bigeyepro-agent
-
Instrument your application
- Use Big:eye Pro SDKs or exporters for your language/framework (Node.js, Python, Java, Go). Example Node.js:
bash
npm install @bigeyepro/sdkjs
const bigeye = require(’@bigeyepro/sdk’); bigeye.init({ apiKey: process.env.BIGEYE_API_KEY, serviceName: ‘my-service’ }); - Add tracing spans and custom metrics where useful (DB calls, external requests, heavy loops).
- Use Big:eye Pro SDKs or exporters for your language/framework (Node.js, Python, Java, Go). Example Node.js:
-
Configure integrations
- Enable integrations for databases, message queues, cloud providers, and web servers from the console or agent config (e.g., PostgreSQL, Redis, AWS CloudWatch).
- Verify metrics/logs/traces appear within 2–5 minutes.
Core UI workflows
- Dashboards: Use default dashboards for CPU, memory, latency, error rate. Create custom dashboards by combining metrics and traces.
- Alerts: Set alert rules on thresholds (e.g., 90th-percentile latency > 500ms for 5m). Configure notification channels (email, Slack, PagerDuty).
- Trace search: Filter traces by service, operation, status code, or duration to find slow requests.
- Log correlation: Link logs to traces for root-cause analysis — open a trace span and view associated log lines.
Practical tips & best practices
- Start small: Monitor a few critical services and the most important metrics first (latency, error rate, CPU, memory). Expand gradually.
- Use percentiles, not averages: Track p50/p95/p99 for latency to surface tail latency issues.
- Tag consistently: Add environment, service, and role tags to metrics and logs to enable focused dashboards and alerts.
- Alert on symptoms, not causes: Alert on user-impacting metrics (error rate, availability, high latency) rather than internal counters alone.
- Reduce noise: Use multi-condition alerts (e.g., high CPU + high load) and mute flapping alerts during deploy windows.
- Instrument key flows: Add tracing to authentication, payment, and heavy DB operations for targeted troubleshooting.
- Retention strategy: Keep high-resolution metrics for critical services longer; downsample or aggregate lower-value metrics to reduce cost.
- Secure keys and access: Rotate API keys periodically and use least-privilege roles for team members.
Troubleshooting common issues
- Agent not reporting: check agent logs (/var/log/bigeyepro/agent.log), validate API key, confirm outbound network to Big:eye Pro endpoints.
- Missing traces: ensure SDK initialized before service start, confirm sampling settings (increase sampling for dev).
- Alert spam: raise thresholds, increase evaluation window, use grouping and suppression rules.
Example checklist for first 24 hours
- Create project and generate API key.
- Install agent on one host and verify it registers.
- Instrument a sample app and confirm traces appear.
- Import or create a dashboard for key metrics.
- Create one critical alert (e.g., error rate > 5% for 5m to Slack).
- Run a load test or simulate errors and verify detection and notification.
Where to go next
- Expand integrations to databases and cloud services.
- Create role-based dashboards for SREs, developers, and managers.
- Automate alerts and runbooks for common incidents.
If you want, I can create a step-by-step install script for your OS (macOS, Ubuntu, RHEL) or an example instrumentation snippet for a specific language — tell me which OS or language.
Leave a Reply