Dashboard Page

Dashboard Page

Dashboard page gives you a summarized view of your Lambda functions in a project at a glance. Displaying metrics and performance information on dashboard page allows you to observe health of your serverless system. You can easily detect problems in Lambda functions with anomaly detection for invocation, error counts and duration of functions.

Dashboard page includes following sections:

  • Statistics

  • Alerts & Insights

  • Lists of Functions

  • Charts with anomalies

Statistics

In Statistics section, general metrics about Lambda functions are displayed. These metrics are calculated according to the selected project and time range. Following metrics are displayed in Statistics:

  • Invocation Count - Total invocation count all of Lambda functions in selected project within selected time range.

  • Error Count - Total error count in invocations for the selected project within selected time range.

  • Cold Start Count - Total number of cold starts of invocations in the selected project within selected time range.

  • Timeout Count - Total number of timeouts of invocations in the selected project within the selected time range.

  • Total Cost - Overall cost of functions in the selected project within the selected time range

  • Health - Overall health of stack for the selected project within the selected time range.

Alerts & Insights

Alerts and insights section gives you an overview about your serverless stack. If you have any created alert policies on Thundra and any of them is violating, you can display them here.

Thundra also gives users insights to discover potential problems on their system. Users can display insgihts for following issues:

  • If there is a jump on invocation count of a function more then 50% compared with the previous period,

  • If any function reaches it's timeout limit of 90% within selected period,

  • If any function reaches it's memory limit of 90% within selected period,

  • If there is a jump on duration of a function more than 50% compared with the previous period.

Users can click on an insight to display details of that function for selected time range.

Charts with Anomalies

Two different charts are displayed on Dashboard page to illustrate invocation, error and duration metrics of your functions. Also, Thundra provides you to observe anomalies on your serverless stack using these charts. Details of how to use this feature explained below.

Anomaly Detection

Thundra finds patterns on your invocation data and find any anomalies within selected time range. On dashboard page, invocation-error count and invocation duration graphs visualize this pattern and emphasize anomalies.

To enable/disable anomaly detection on the invocation&error count chart, click on button Anomaly.

You can select two parameters to find pattern and detect anomalies from gear button on the right corner of the chart section. This selections affect both charts.

  • Period - Time range of recurring pattern of data.

  • Rollup - Time interval to aggregate data points to one point. If you set up rollup to5 minutes, all data in 5 minutes will be displayed as one data point.

Chart with Anomalies

Grey area represents normal pattern of your data. If any data point has value below or above of this pattern it will be displayed as red. When you hover on data points, you can display boundaries of normal pattern and value of anomaly data.

Invocation and Error Count

Invocation&Error Count Graph

Invocation, error and cold start count changes within the selected time range are visualized in this graph. Users can select which metrics will be displayed on chart using buttons above. One of the metrics should be selected. If anomaly detection is on and any anomalies detected, they will be displayed with red points. Following statistics are displayed when hovered on chart:

  • Invocation count

  • Error count

  • Cold start count

Statistic&Anomaly selection

If anomaly is on, for each selected statistic following values will be displayed when hovered:

  • Anomaly bound - Boundary values (min-max) for normal instances

  • Anomaly detected - Value of detected anomaly

Users can select a function from left table to display error and invocation counts on the chart. Click on function name at the top of the chart to navigate to detail of it.

Invocation Duration

Invocation Duration Chart

Following data is displayed when hovered on invocation duration chart:

  • Average duration expected - It will be displayed if anomaly is enabled.

  • Average duration anomaly - Average duration that outside of normal patterns. It will be displayed if anomaly is enabled.

  • Average duration

List of Functions

Most Costly Functions

To give an insight about functions in user's stack, two different function lists are displayed on Dashboard page. Users can select a list from dropdown above functions table. Up to 15 functions can be listed.

Function List Dropdown
  • Most costly functions - Functions that have the most cost are listed.

  • Most problematic functions - Functions that have the worst health are listed. Metrics such as error count and timeout count affect health of a function.

Most Problematic Functions

If you want to have a deeper insight about a function, click on it. Charts that are displayed in right hand side will be re-drawn according to selected function. Using this feature, you can easily observe metrics and anomalies of your function. Click on function name at the top of charts to navigate to details of a function. To display charts for all function in project, click on close icon on the right corner.