Prometheus Exporter Pro - Using Dashboard Feature

We are excited to share that you can now effortlessly monitor the health status of your data center directly from any instance within a cluster using the dashboard feature. This addition is a significant step forward, offering a first-stop solution for data analysis and visualization. Previously, customers might have relied on third-party tools like Grafana for data visualization. While Grafana offers extensive functionalities, our integrated dashboard simplifies the process, providing immediate and intuitive access to crucial metrics right within your Atlassian platform.

There are four key tabs for providing overall status:

• JVM (all JVM metrics)

• Request Duration On Path

• Instance Specific(plugin manipulations, user’s info, repository updates per node)

• Table representation (JMX metrics and other JVM data)

JVM Metrics

This tab provides a comprehensive view of Java Virtual Machine (JVM) metrics, crucial for assessing the health and performance of the system. It includes:

• Memory Usage: Detailed insights into heap and non-heap memory usage, helping in identifying memory leaks or areas of high memory consumption.

• Garbage Collection: Information on garbage collection events, including frequency and duration, aiding in tuning JVM for optimal performance.

• Thread Count: Monitoring of thread count and state, crucial for diagnosing concurrency issues or deadlocks.

Open

Request Duration on Path

This tab focuses on monitoring the performance of HTTP requests:

• Path-wise Breakdown: Performance metrics for each HTTP request path across the nodes in your cluster, aiding in identifying slow-performing endpoints.

• Response Time Analysis: Average, minimum, and maximum response times for requests, useful for performance tuning and capacity planning.

Open

Instance Specific

Targeted at monitoring individual nodes in a cluster, this tab includes:

• Plugin Manipulations: Tracks changes or updates to plugins, ensuring they are functioning as intended.

• User Information: Monitors user activities, providing insights into system usage patterns.

• Repository Updates: Keeps track of repository changes per node in real-time, crucial for maintaining data integrity and consistency.

Table Representation

Presents JMX metrics and JVM data in a tabular format:

• Metric Details: Each row represents a specific metric with detailed statistics.

• Customization: Allows users to customize which metrics are displayed for focused monitoring.

Use Cases and Applications for Bitbucket Platform

1. Optimizing Bitbucket Server Performance

A software development company relies heavily on Bitbucket for its version control and collaborative needs. However, they started noticing slow pull request operations and overall sluggishness in Bitbucket server response. By utilizing the JVM Metrics tab, the IT team could monitor the JVM performance of their Bitbucket servers. They identified a pattern of high heap memory usage during peak hours, indicating the need for JVM tuning and possibly scaling up resources. This led to a significant improvement in Bitbucket server performance, ensuring smooth operations during high-traffic periods.

2. Analyzing HTTP Request Patterns

In another instance, a large enterprise using Bitbucket for source code management observed inconsistencies in HTTP request performance, particularly in repository access and pull request operations. By employing the Request Duration on Path tab, the DevOps team could break down the response times for various HTTP paths. They discovered certain API endpoints related to pull requests were consistently underperforming. This insight allowed them to focus their optimization efforts on specific areas of the Bitbucket server, leading to enhanced efficiency and user experience.

3. Managing Plugins and User Activities

For a multinational corporation with a distributed development team, managing Bitbucket plugins and user activities was becoming a challenge. The Instance Specific tab provided a comprehensive view of plugin manipulations and user activities across different Bitbucket nodes. This enabled the IT administrators to efficiently manage plugin updates, ensure balanced distribution of repository activities, and maintain high system performance across all nodes.

4. Detailed JMX Metrics Analysis for Troubleshooting

A tech startup faced issues with its Bitbucket instance crashing intermittently. Using the Table Representation tab, their technical team could delve into detailed JMX metrics and JVM data, uncovering issues with garbage collection processes and thread management. This granular level of data helped them to pinpoint the root causes of the crashes and implement specific fixes to stabilize their Bitbucket environment.

Troubleshooting and FAQs

Q1: How often is the dashboard data refreshed?

• A: It’s configurable on the Prom Exporter Configuration page by setting the job interval(duration parameter). App updates this data during metrics collection in a background job.

Q2: Can I customize what metrics are shown in the Table Representation?

• A: We plan to add thresholds for our embedded alerting system. At this moment we can walk through all the available metrics parameters on predefined flexible charts without customization.

Q3: Can Prometheus Exporter Pro monitor multiple Atlassian products simultaneously?

• A: It shows metrics from all nodes in a cluster(but it doesn’t combine data from different applications, e.g. Confluence + Jira, etc..)

Q4: Is there a mobile version or app for Prometheus Exporter Pro for on-the-go monitoring?

• A: Currently, Prometheus Exporter Pro functions as a plugin that is installed directly on Atlassian platforms like Jira, Confluence, or Bitbucket. It does not have a separate mobile app for on-the-go monitoring.

Q5: How does Prometheus Exporter Pro help in identifying Bitbucket performance bottlenecks?

• A: The tool provides detailed JVM metrics and HTTP request analysis, which helps in pinpointing performance issues such as slow pull requests or high server load times in Bitbucket.