Apivom Otto — metrics and alerting for the whole platform
Otto is the metrics and alerting layer behind the Apivom products. It collects numbers from hosts and services every 15 seconds, checks that each service answers, and emails operators when something is wrong. Otto is included with the products and is not sold separately.
Infrastructure layerOtto in numbers
Know the state of the platform, and hear about trouble first
Every host and service reports to one place, so a slow disk, a stopped service or a spike in errors turns into an alert instead of a surprise.
Collected every 15 seconds
Metrics from hosts, databases and application services are collected on a fixed 15-second interval, so the picture is always recent.
Time series, not snapshots
Readings are kept as time series, so engineers can follow a trend and compare a problem with how the same service behaved before.
Health checks for each service
Otto probes the health endpoint of the Apivom apps and platform services, so a service that stops answering is noticed at once.
Alert rules with severities
Rules cover availability, host load, memory and disk, the web servers and Odin’s ingestion. Each alert carries a severity.
Repeats until resolved
Critical alerts repeat every hour, warnings every 4 hours and info every 24 hours, and a resolved notice is sent when the problem clears.
Feeds the dashboards
Otto is the data source for Apivom Ping and for the service status panels in the admin console.
How the Apivom products use Otto
You do not set Otto up on its own. It watches the hosts and services the Apivom products run on, and Apivom operators use it to see their state. Here is what it covers today.
- Hosts and datastores. CPU, memory, disk, network and service units of the host, plus the databases and the cache.
- Application services. Halo’s voice server, Orbit’s event pipeline, the document converter, Odin, the Staff cluster and Blob.
- Health checks across the apps. Atlas, Iris, Pivot, Mars, Orbit, Abu, Halo, Staff, Vesta, Key, Blob, Mask, Opus, Ping, Odin and the admin console and websites.
- Custom metrics from timers. Scheduled jobs can publish their own numbers. Odin’s canary and the web servers’ error counters use this.
Alerts that do not go quiet
An alert that is sent once is easy to miss. Otto keeps repeating by severity until the problem is resolved, and tells operators when it is.
- Rules for what matters. Services down or flapping, host CPU, memory, disk, load and reboots, platform and edge health, web server errors and Odin ingestion.
- Email to a shared mailbox. Alerts reach operators by email through the platform mail service, including resolved notices.
- Sign-in through Key. Operators sign in through Key with the Otto access role. Health paths stay open for edge probes.
Frequently asked questions
What is Apivom Otto?
Otto is metrics and alerting for the Apivom platform. It collects metrics from hosts and services every 15 seconds, keeps them as time series, probes service health endpoints, evaluates alert rules with severities and emails operators until a problem is resolved.
Is Otto sold separately?
No. Otto is an infrastructure layer that is included with the Apivom products. You do not buy or set it up on its own. If you want to know how it supports your use of the products, talk to us.
What does Otto measure?
Host CPU, memory, disk, network and service units, the databases and the cache, Halo’s voice server, Orbit’s event pipeline, the document converter, Odin, the Staff cluster and Blob. It also checks the health of the Apivom apps and platform services.
How do alerts work?
Rules watch availability, host load, memory and disk, platform and edge health, web server errors and Odin’s ingestion. Each alert has a severity and goes out by email. Critical alerts repeat every hour, warnings every 4 hours and info every 24 hours until resolved.
How does Otto relate to Apivom Ping and Apivom Odin?
Otto collects the numbers and raises alerts. Ping is the dashboard layer that reads them. Odin tracks errors and traces, and Otto watches that Odin itself is ingesting events, so silence does not go unnoticed.
Can customers see Otto?
Otto is a tool for Apivom operators, who sign in through Key. Customer-facing dashboards and status pages are not part of what it offers today. If you need to know the state of something you use, talk to us.
Trouble should reach us before it reaches you
Otto comes with the Apivom products. Talk to us about how Apivom keeps an eye on the platform your work runs on.