Observe

Observe stores your logs, metrics, and traces as datasets, queried with OPAL, Observe’s pipeline language. Nexus queries it to see what your services were doing around the time of an incident: the log lines, the metric that moved, and the slow or failing span behind a bad request.

​

What we support

Connecting Observe gives Nexus three data sources, which you can enable and disable independently:

​

Logs

Nexus filters and searches your log datasets to read what a service was logging at the time of an incident. OPAL can also aggregate log lines, so an error count climbing or a request volume dropping away shows up as a trend even where you never set up a dedicated metric for it. It filters on the fields your logs actually contain.

​

Metrics

Nexus queries your metrics and graphs them for the incident’s time window, so a resource spike, a latency change, or a growing error count shows up against the period that matters. It can narrow a metric to the service or host involved.

​

Traces

Nexus uses your traces in two ways: searching for the spans behind a problem (a slow endpoint, a failing dependency, a particular service) without already knowing a trace ID, and pulling a specific trace by ID to see the full span tree and where it broke.

​

Your datasets, discovered rather than assumed

In Observe, the datasets you have depend on the apps you’ve installed and how you send data. Tracing, Host Explorer, and Kubernetes Explorer each build their own curated datasets, and data you send over OpenTelemetry without those apps lands in a raw datastream. No two Observe accounts are laid out the same way, so Nexus discovers your datasets and the fields in them instead of expecting fixed names. It reads curated datasets and raw datastreams, so it can still query your data before you’ve installed any apps. When part of the incident window isn’t available in a dataset yet, Observe returns a partial result, and Nexus treats it as partial. A sparse result then isn’t mistaken for a quiet period. Nexus learns this structure automatically. How that works is covered in How telemetry works.

​

Limitations

• No dashboards. Nexus doesn’t read your Observe dashboards.
• Links go to your Observe account, not to a query. When Nexus cites an Observe query, the link opens your Observe account rather than the query it ran.
• Large results are cut off. Observe returns at most 100,000 rows for a query. Nexus marks a result that reaches that limit as truncated, rather than treating it as complete.
• Raw datastreams hold every signal. A raw datastream holds logs, metrics, and spans together, and each Observe data source can read it. Disabling one data source doesn’t stop the others from reading that signal’s rows in a raw datastream.

​

Connecting Observe

You connect Observe directly. There’s no provider in front of it. What you’ll need:

• Your Observe customer ID: the number at the start of your Observe URL, such as 123456789012 in https://123456789012.observeinc.com.
• Your Observe region: the domain after your customer ID in that URL. We support US West and US Central (observeinc.com), US East (us-2.observeinc.com), Canada (ca-1.observeinc.com), Frankfurt (eu-1.observeinc.com), United Kingdom (eu-2.observeinc.com), Tokyo (ap-1.observeinc.com), and Sydney (au-1.observeinc.com). Observe lists these on its regions page.
• An Observe API token for a service account. In Observe, create a service account under Manage account → Service accounts, give it view access to your logs, metrics, and tracing datasets, and create an API token for it. See Observe’s guide to API tokens.

1. From the Nexus telemetry settings, add a telemetry data source and choose Observe.
2. Enter your customer ID, choose your region, and paste the API token, then test the connection. We check that Observe accepts the token in that region and that the token can see your datasets.
3. Save. Nexus adds Observe with logs, metrics, and traces all enabled. To stop Nexus using one of them, disable it from the data source’s page.

​

Rotating your API token

Observe service account tokens expire, and using a token doesn’t extend it. The connection test warns you when your token has less than a week left. To replace it, open the Observe data source in the Nexus telemetry settings and update the token. We test the new token before saving it. Your customer ID and region can’t change after you connect.

​

Best practices

• Use a token that belongs to a service account rather than to a person, so it keeps working if they leave.
• Give the service account view access only. Nexus only reads from Observe.
• Set yourself a reminder before the token expires, and rotate it before then.
• Enable logs, metrics, and traces together where your team uses them. With more than one connected, an investigation can move from a log line to the trace behind it, or from a metric to the logs explaining it.

​

Related