OGC OWS Monitoring Challenges

“I see pink tiles!”

What we expect

What we see

What we received

In Image Error Message

"The Blank Stare"

But our HTTP Monitor said:

200 OK...

GetCapabilities response OK, but

• Capabilities doc may be static file
• No guarantee specific requests will work (WMS GetMap, WFS GetFeature, etc.)

Time-based OGC Services

SensorWeb Enablement (SWE)

• Internet of Things (IoT)
• Sensor Observation Service (SOS)
• SensorThings API (STA)

SOS Viewer (52North)

Gaps in SOS-Data

Public "Uptime" services

• Generic HTTP checking (keywords)
• OWS specifics need deeper inspection
• Not accessible: most critical OGC-services run internally on intranets

Value proposition

• OWS Aware Web Service Monitor
• Quality of Service (QoS) Checking
• with History Capture

GeoHealthCheck Overview

and Tour

Overview

• Founded by Tom Kralidis
• Started in the air, literally!
• In flight en route to FOSS4G 2014 (YYZ -> YYC -> PDX)

A geopython Project

Tour

demo.geohealthcheck.org

Failing Resource!

Root Cause Analysis

Architecture

GHC Parts

• Python Webapp (Dashboard)
• HealthCheck Runner (Plugins!)
• Database

Dashboard

• Configure/Manage Healthchecks
• Flask Webapp
• Standard Python WSGI
• Run with Nginx, Apache2, Gunicorn...

HealthCheck Runner

• Runs Healthchecks (duh!)
• Driven by Scheduler (APScheduler)
• Reporting (Down/Up, Daily Status)
• Notifications: email, webhook, ...

HealthCheck Model

• Resource: abstraction for (OWS or REST) API Endpoint
• Resource: has (API endpoint) URL
• Probe: fires request(s) on (Resource) URL
• Resource has 1-N Probes

HealthCheck Model

• Probe has N Checks (checklist)
• Each Check: evaluates Probe result aspect (e.g. response content)
• Check gives boolean verdict (success/fail)
• All Probe Checks: Probe Run Report (JSON)

Plugins - Probes and Checks

• Standard (included in GHC)
• Custom (include your own!)
• Configurable via Web UI
• More later on...

Database

• Entities
  • Users, Resources, Runs, Tags
  • Probes, Checks, Recipients
• Per-Probe run schedule
• Maintains history (Runs)
• Multiple backends (via SQLAlchemy: default SQLite)

Tags

• For grouping Resources
• Provide/manage in Dashboard

Data Model

Installation

Standard

• Standard Python setup - Instructions
• Paver for setup and management tasks
• Alembic with Flask-Migrate (for DB upgrades)

Use Docker!

• Versioned GHC Images on DockerHub
• Docker Compose support: (complete stack: Webapp, DB and Jobs)
• Documentation
• GHC Up & Running in minutes!!

Settings

							
GHC_RETENTION_DAYS = 30
GHC_RUN_FREQUENCY = 'hourly'
GHC_SELF_REGISTER = False
GHC_NOTIFICATIONS = False
GHC_NOTIFICATIONS_VERBOSITY = True
GHC_ADMIN_EMAIL = 'you@example.com'
GHC_NOTIFICATIONS_EMAIL = ['you2@example.com', ..]
GHC_SITE_TITLE = 'GeoHealthCheck Demonstration'
GHC_SITE_URL = 'http://host'
GHC_SMTP = (email settings)
GHC_RELIABILITY_MATRIX = (when to show green/orange/red)
GHC_MAP = (map setup)
GHC_PLUGINS = (Built-in Probes and Checks)
GHC_USER_PLUGINS = (Probes and Checks, YOURS!)
GHC_PROBE_DEFAULTS = (Default Probe per Resource Type)
GHC_REQUIRE_WEBAPP_AUTH = False
							
						

Full list of configuration parameters

Extending GHC

Plugin Development

Plugin Model

• Probes and Checks are Plugins
• Plugin class and/or modules in config
  • Built-in Plugins: GHC_PLUGINS=
  • Your Plugins: GHC_USER_PLUGINS=
• Must be in $PYTHONPATH

Plugin Model

Two Probe Types

• Template (OWS) Requests
• Free-form: Probe Anything!

Time for some code!

See also Plugin Docs

Simplest Probe Class

Check Class HTTP Status

Base GetCapabilities Probe

WMS GetCapabilities Probe

Roadmap

Planned

• REST API architecture
• UI renewal (Vue.js)
• Monitoring tools integration (Icinga, Munin, Prometheus etc.)
• OGC Quality of Service and Experience Domain Working Group (QoSE DWG)

You Can Help!

• Coding (Plugins!)
• Testing
• Documentation
• User Stories
• Sponsored Development
• Translation

Support

Shameless Plug

Cloud Subscription

geoqos.com