Is analytics suitable for mission critical data?

Not applicable

We plan to store custom attributes in Apigee analytics and use them to make business decisions, so it is important that we don't loose data.

We have the following questions:

1. What is the level of confidence on the storage of data in analytics? Are there scenarios in which data might be lost?

2. How does analytics handle performance? For example, if we have a million transactions every day and we want to fire "group by" queries on the 5-10 custom fields we have.

0 3 245
3 REPLIES 3

Analytics is a critical component; the confidence must be high for it to be valuable. That said, I think it's always important to think about risk mitigation, so I would retrieve the relevant data via API and put it into a backup solution.

Analytics performance will not impact API traffic, so the main concern is how quickly can the data be retrieved. Edge Analytics stores aggregates already, so the number of transactions per day is not as significant as the amount of data retention.

Not applicable

Analytics does support custom queries, however the current implementation does not optimize schema to support those queries. Your custom attributes will not be indexed, and unless you configure them in precisely the correct way you may not be able to aggregate across multiple proxies with those attributes.

Apigee does support custom retention periods - you can work with support on ensuring you have the right package in place for that.

In the end, I would recommend you look at what @Carlos Eberhardt is recommending - a secondary data warehouse where you can optimize queries and manage the lifecycle of data.

Thanks David. Could you elaborate on "unless you configure them in precisely the correct way you may not be able to aggregate across multiple proxies with those attributes"?