Monitoring

Monitoring

In this guide:

• View latency and errors in Grafana

• View service logs in CloudWatch

View latency and errors in Grafana

Access the Grafana dashboards here. Dashboards are accessible from the sidebar menu from the home page. Up to 2 days of data is available.

Note: The 2 day data limit is a cost savings measure due to very high TPS. Storing more logs for longer periods of time would make FireFly unsustainable.

You’ll see folders containing dashboards. The key ones are:

• Firefly - Clients and Services

• Firefly - Platform

For example, if you’re looking for metrics on FireFly’s requests to your service, see the Services Overview dashboard.

Here you can view latency, TPS, and error rates. You can also filter by the specific API, region, or percentile.

Best practices

• Do not move graphs

• Discard and do not save any dashboard changes

Onboarding a new service

If you have onboarded a new service and want to see it in Grafana, create a trouble ticket.

Alarms

• If your service is not performing, the Devex API team may page you.

• If your service needs to be alarmed:

  • Determine your service, latency, threshold and CTI numbers.

  • Write a trouble ticket and the Devex API team will create the alarms for you.

Logs

Gaining Access

Log data is highly sensitive and access is carefully protected. If this is your first time accessing production logs, it’s possible that we have not onboarded your team yet.

If you are not onboarded, go to the Bindle page to request that the Devex-API team grant developer permissions to yourself or your team:

1. In the dropdown, select the role below Conduit IAM Roles

2. Click Console Access

3. In the following popup window, click Console Access a second time

4. In the next popup, under Description write in the reason for your need for access

5. Click Submit Contingent Authorization

Note: if you pause during the process, you should leave the page open, otherwise your request will be canceled. The prompt will expire in 2 hours.

• Confirm your access on this Bindle page.

Searching the Logs

1. Sign into the AWS account (305080342773) using View Only access.

2. Navigate to the CloudWatch console, in the same region of the request you are finding logs for.

3. Navigate to the appropriate next location from the sidebar.

   • For simple filtering, select Log groups.

   • For more complex queries, select Logs Insights.

4. Choose the log groups graphql and graphqlCore. Your request may be in graphqlCore logs if it the request’s region does not match the account’s default region.

5. Filter down to the logs you’re searching for, using the appropriate time range and keywords.

   • If using Log groups, try entering a keyword like requestId, customerId, or an id from your query

   • If using Logs Insights, see the CloudWatch Insights Cookbook for help with creating more complex queries.