Getting started video
An 8-minute dashboard walkthrough covering account setup, two-factor authentication, incident analysis, code and pull request scans, metrics, APM, and agents.
Watch the walkthrough
Follow the full tutorial or use the chapter buttons to jump to the task you need. The video includes British English narration and captions.
Jump to a chapter
Follow along
Open your Reelay workspace alongside the video. Start with account creation and authenticator setup, then explore incidents, errors, scans, metrics, and agent health. When you are ready to connect your own applications, follow the Quickstart and SDK guides.
Video transcript
Read the video transcript
Create an account
Welcome to Reelay. In this walkthrough, we will create an account, set up two factor authentication, review incidents and errors, run code and pull request scans, and check metrics and agents. The workspace and results shown here are sample data.
On the signup page, enter your organisation name and work email. Choose a password with at least ten characters, including uppercase and lowercase letters, a number, and a special character. When all the requirements are satisfied, select Create workspace.
Next, check your inbox for the six digit verification code. Enter the code into these boxes and select Verify and continue. If the email has not arrived, check your spam folder and use the resend option when needed.
Secure your account
The next screen secures your account with an authenticator. Open your authenticator app on your phone, add a new account, and scan this QR code. If scanning is unavailable, enter the setup key manually. This tutorial uses a demonstration key.
Your authenticator now generates a changing six digit code. Enter the current code here, then select the confirmation button. This code comes from the authenticator, rather than your email. If it expires while you are typing, use the next code.
Incident analysis
After setup, open Incidents from the sidebar. This list is where you track operational issues. Use search and the available filters to narrow the list. Open the incident you want to investigate, and check its severity, status, and affected node.
Inside the incident, start with the Diagnosis tab. Read the incident context and customer impact before looking at the analysis. In our example, checkout requests are failing after a deployment. The sample analysis connects those failures to the checkout calculation.
Read the proposed root cause and explanation, then inspect the suspect files. Use the reported location to compare the finding with your code and deployed revision. Treat the analysis as an investigation aid, and confirm it against the available runtime evidence.
Scroll down to the suggested fixes. Here, the recommendation is to validate the checkout payload before reading the total. Review the proposed change, reproduce the failure where possible, and test the correction before releasing it to your customers.
Use Timeline to follow the sequence of updates and response activity. This helps your team understand what happened and what has already been checked. Keep the incident status and response notes current as the investigation moves towards resolution.
Errors and scanning
Open Errors in the sidebar to review application failures captured by your instrumentation. Search for the relevant message or narrow the list with filters. Look at recurring errors and their incident links to decide which failure needs attention first.
Open an error to inspect its message and captured event details. The stack trace helps you locate the failing function and line. Compare the environment and request context with the affected service, then follow the linked incident when coordinated investigation is needed.
For repository scanning, open Codebase Analysis. This is where you review scan history and code findings. Runtime errors and repository scans provide different evidence, so use both during debugging. To start a new repository check, select Quick scan.
The quick scan form asks for an authorised source control connection and repository. If your repository is unavailable, connect it through Integrations first. Choose the repository you want to inspect, rather than starting with an unrelated service or branch.
Select the branch to scan and add a name if you want to identify this run later. In this example, we scan the main branch of the checkout repository. Check the source carefully before moving on to the project checks.
Choose the language or framework preset, then review the install, build, and lint commands. These commands must match your repository scripts. Tests are optional in this form. OpenGrep and Trivy run on every scan, and AI review is an optional additional check.
Once the repository, branch, and required commands are ready, select Start quick scan. In a live workspace, the run is queued and progresses through its checks. Wait for completion before interpreting the results; a running scan does not yet provide the final outcome.
When the run completes, review its decision and findings. Use severity filters to focus your review, then open the finding for its file, line, explanation, and suggested correction. Check the evidence and fix the underlying issue before running the checks again.
Pull request scanning
Open PR Scans to review pull request checks and monitors. The page shows monitor status and recent run history. Open a monitor to inspect its runs, or select Scan a pull request to configure a comparison for a specific branch.
Choose your repository, then select the source branch and target branch. Here, the checkout validation branch is compared with main. Review the project commands and optional AI review setting before selecting Scan pull request. Make sure the comparison points in the intended direction.
Use New monitor when you want repeatable pull request checking. Select the repository and target branch, then configure commands that work for this project. Give the monitor a useful name. After creation, return to its history to review future scan results.
Metrics and APM
Select APM slash Metrics in the sidebar. This page brings application performance measurements together for your nodes. Your application must send APM data before useful charts can appear. An empty chart should prompt you to check instrumentation and recent traffic.
Choose the affected node and the time window you want to investigate. Start with the period around the incident, then widen the window for context. Comparing the same node and time range helps you connect a change in errors with a change in performance.
Review transactions and throughput together. Throughput here is measured per hour. Next, check the error rate and latency charts. P ninety five latency means ninety five percent of requests were faster than that value. Use the latency selector to compare percentiles. Scroll down to Apdex to review the satisfaction score alongside these measurements.
Agents
Open Agents to manage agent access and inspect connected deployments. The available setup controls depend on your workspace entitlement and configuration. Agent tokens authenticate deployments in your infrastructure. Use a descriptive token name so your team can identify its purpose.
In Connected agents, review each hostname, version, status, and latest check in. Use search to find a particular deployment. A recent heartbeat is important: a registered agent can still be offline. Refresh this page after starting or updating an agent.
Use the Offline filter to focus on agents that are not currently checking in. Check the deployment process, its configuration, network access, and logs. After correcting the issue, refresh the list and confirm that the agent has returned to an online state.
For a new agent, follow the deployment guide on this page. Use the image and environment settings provided for your workspace, and store credentials securely. After deployment, return to Connected agents and confirm a fresh heartbeat before relying on its data.
Ready to use Reelay
You are ready to get started. Secure your account, review incidents and errors, inspect code and pull request findings, compare your performance charts, and keep agents online. Return to the relevant node and time window whenever you investigate a new issue. Thank you for watching.