September 16, 2019 By Glynn Bird 4 min read

Cloudant is now rolling out customer-facing logging for its Cloudant services.

In this post, we’ll explore how to set it up and how it can help you keep tabs on your IBM Cloudant service.

Note: At the time of writing, this service is only available on our Dallas, London, and Frankfurt-based cloud clusters. Other services will follow in due course. Setting up logging

First, we need to set up an IBM Log Analysis with LogDNA service in the IBM Cloud. Begin by choosing the service from the Developer Tools section of the IBM Cloud catalog:

Configure the service in the same region as your Cloudant service (e.g., Dallas) and choose and appropriate a resource group:

Once provisioned, choose the Configure Platform Services button to direct your IBM Cloud platform logs to your LogDNA service:

That’s it! Click the View LogDNA button to see a live stream of your logs:

You should see all of your API calls listed in the live stream, with a short delay as the data makes its way to LogDNA.

Logging data structure

The data is logged as a JSON object containing everything you need to know about the request and how it was handled:

{
    "accountName": "mycloudantaccount",
    "httpMethod": "GET",
    "httpRequest": "/cities/_all_docs?limit=10",
    "responseSizeBytes": 1221,
    "clientIp": "94.194.94.13",
    "clientPort": 63072,
    "statusCode": 200,
    "terminationState": "----",
    "dbName": "cities",
    "dbRequest": "_all_docs",
    "userAgent": "curl/7.54.0",
    "sslVersion": "TLSv1.2",
    "requestClass": "query",
    "parsedQueryString": {
        "limit": "10"
    },
    "rawQueryString": "limit=10",
    "timings": {
        "connect": 142,
        "request": 0,
        "response": 10,
        "transfer": 1
    },
    "meta": {},
    "logSourceCRN": "crn:v1:bluemix:public:cloudantnosqldb:us-south:a/someid::",
    "saveServiceCopy": false
}

In the free version of LogDNA, only the live stream of data can be viewed, with optional filtering by source. In the paid plans, the JSON data is parsed and is retained for a period of time, with optional archiving in IBM Cloud Object Storage.

With the data parsed, it’s easy to compose queries, alerts, and graphs based on the data you see in the object above. You may wish to keep an eye on the following:

  • The response timing of query requests on a particular database
  • The count of the occurrence of HTTP 429 responses, indicating that you have exceed your provisioned throughput
  • The range of IP addresses that are accessing your Cloudant service
  • The number of requests hitting each of your databases
  • The version TLS being used by your client code to access client: Do you have traffic originating from an out-of-date source?
  • The values of responseSizeBytes that are very high: Do some of your requests return lots of data?

The LogDNA documentation shows how you can explore the incoming stream of data, creating named views which can be retrieved later. Saved views can be combined to create visual dashboards to give you an at-a-glance view of your Cloudant traffic.

Example graphs

Request counts by HTTP request code, grouped into successes (statusCode < 300) and failures (statusCode >= 400):

Request counts by request “class” (i.e., whether IBM considers the request a read, write, or query for billing purposes):

Request counts by HTTP method (get or post, in this case):

95th percentile response times broken out by each of the timing metrics:

Creating a graph—step by step

1. Log into your Log Analysis with LogDNA service, choose Boards from the left-hand menu and click New Board.

2. Click Add Graph.

3. LogDNA then needs to know which field in the JSON you wish to plot:

4. Enter statusCode, the < operator, and 300. This will only show status codes of 299 and below, indicating success:

5. You should now see a plot of the counts of successful API calls against time. To add a second plot, click Add Plot:

6. This time, choose statusCode >= 300 to plot the errors:

7. The graph will change to add a second plot, showing the error counts alongside the success counts.

8. Repeat this process to add further graphs on metrics that interest you.

Timing metrics

The logged JSON contains an object with the timings of each part of the request:

"timings": {
  "connect": 142,
  "request": 0,
  "response": 10,
  "transfer": 1
}
  • connect: The time it took for your HTTP client to make the initial HTTPS connection (the handshakes required to negotiate a secure connection between client and server). Programmatic API calls using the official Cloudant libraries will reuse existing connections, eliminating this connection delay.
  • request: The time taken to receive the request data. For bulk uploads, this may be significant, especially from clients with limited uplink bandwidth.
  • response: The time taken for the database to do its work.
  • transfer: The time taken for the response to be transferred to the client.

Further information

More from Announcements

Unify and share data across Netezza and watsonx.data for new generative AI applications

3 min read - In today's data and AI-driven world, organizations are generating vast amounts of data from various sources. The ability to extract value from AI initiatives relies heavily on the availability and quality of an enterprise's underlying data. In order to unlock the full potential of data for AI, organizations must be able to effectively navigate their complex IT landscapes across the hybrid cloud.   At this year’s IBM Think conference in Boston, we announced the new capabilities of IBM watsonx.data, an open…

IBM and SAP unlock business and industry value with new generative AI solutions 

3 min read - IBM Consulting is delivering on our commitment to co-innovate with SAP and collaborate with our clients. As part of our Value Generation Partnership initiative announced earlier this month with SAP, we are releasing the first 10 of 100 planned AI solutions to help clients transform their industries, optimize their business processes and successfully deliver their SAP programs.  Delivering AI business and industry innovation at scale  With the recently announced Value Generation Partnership initiative, IBM and SAP are co-innovating intelligent industry…

IBM SevOne 7.0: Reaching application-centric multicloud network observability  

2 min read - As enterprises increasingly rely on network connectivity to support cloud-based applications and remote workers, network managers require new methods to monitor and safeguard connectivity across diverse environments, including corporate networks, software-defined WANs and multiple public cloud providers.   According to the recent EMA Network Megatrends Report, responding network professionals believe that 53% of network outages and performance issues could be prevented with improved network management tools, yet only 9% find it very easy to hire skilled networking personnel. This is why…

IBM Newsletters

Get our newsletters and topic updates that deliver the latest thought leadership and insights on emerging trends.
Subscribe now More newsletters