Is there any limitation for APIGEE and Salesforce Integration (retrieval)

One of the use-case I am look at is:

- Have the batch/scheduler in APIGEE, to run once in a day.

-> Any limitation/issues on having schedules inside APIGEE proxy, say HT.

-> The scheduled frequency/cron is flexible enough to update in Runtime?

- Invoke Salesforce Bulk Query API to retrieve multiple records from multiple objects.

-> Any limitations here in APIGEE, as far as data it can handle?

-> Also references to any similar implementation would be useful.

- Combine all the responses and feed it to an end-point/rest webservice.

Many thanks!!

Solved Solved
0 2 361
1 ACCEPTED SOLUTION

Hi, I spent some time about your questions, and below I will left some answers:

  • Have the batch/scheduler in APIGEE, to run once in a day.
    • I think that proxies are not designed for this proposal, is important the proxy be called by a request, process the request go to target endpoint and process the response from target endpoint. Proxy, in my opinion, don't should start the process for its self. However, if you want this behavior, you can deploy an app in Java/Python/Node and use something library that will do that for you.
  • Any limitation/issues on having schedules inside APIGEE proxy, say HT.
    • As I said before, it is not good idea, but you can do that. Limitations will happen like any proxy called through request/client, limits with requests, timeout, size of the response, etc.
  • The scheduled frequency/cron is flexible enough to update in Runtime?
  • Invoke Salesforce Bulk Query API to retrieve multiple records from multiple objects.
  • Any limitations here in APIGEE, as far as data it can handle?
    • Normally limitations happens when the data retrieved is huge, I think the size of the response body is 5mb per request. To fix that, you can find some solutions like pagination, for example.
  • Also references to any similar implementation would be useful.
  • Combine all the responses and feed it to an end-point/rest webservice.
    • You can use callbacks through apps resources, with this approach you can use Promises, javascript, and merge the response of each request.

I hope that it help you in some way.

Cleison Melo.

View solution in original post

2 REPLIES 2

Hi, I spent some time about your questions, and below I will left some answers:

  • Have the batch/scheduler in APIGEE, to run once in a day.
    • I think that proxies are not designed for this proposal, is important the proxy be called by a request, process the request go to target endpoint and process the response from target endpoint. Proxy, in my opinion, don't should start the process for its self. However, if you want this behavior, you can deploy an app in Java/Python/Node and use something library that will do that for you.
  • Any limitation/issues on having schedules inside APIGEE proxy, say HT.
    • As I said before, it is not good idea, but you can do that. Limitations will happen like any proxy called through request/client, limits with requests, timeout, size of the response, etc.
  • The scheduled frequency/cron is flexible enough to update in Runtime?
  • Invoke Salesforce Bulk Query API to retrieve multiple records from multiple objects.
  • Any limitations here in APIGEE, as far as data it can handle?
    • Normally limitations happens when the data retrieved is huge, I think the size of the response body is 5mb per request. To fix that, you can find some solutions like pagination, for example.
  • Also references to any similar implementation would be useful.
  • Combine all the responses and feed it to an end-point/rest webservice.
    • You can use callbacks through apps resources, with this approach you can use Promises, javascript, and merge the response of each request.

I hope that it help you in some way.

Cleison Melo.

Hi @Cleison Melo

Firstly thanks for taking time to respond.

Yes, I think I can agree with not maintaining scheduler inside the proxy.

I have also looked at the tentative size we may receive from backend, it will be less than 2mb. Should be okay in that case, based on your response.