We have this table. It contains data which will be picked up by an hourly batch job for processing. Five times a day, there are also multiple batch ETL jobs which get data from other source table and transfers it to this table.
Originally, that was it's only function. Now we were told that there will be a new page in our web application that would read data from this table. My worry is that due to the number of batch job read and writes to the table, would it still be feasible to read data from it and show it in our web page? Our web app has about 30,000 users that would be able to access that new page and it is estimated they most users use the web page everyday and they would want to view the contents as quickly as possible (there's an SLA). Due to the number of read and writes to the table from the batch jobs, achieving the desired page response time in the SLA might not be possible
We are thinking twice of creating a read replicate since the hourly batch job that reads from this table updates statuses of records that are processed. There would also be additional synch jobs needed if we would use a read replica.
Where do we draw the line between saying "OK we need a read replica for this" and "No problem we still can have one table to do all that" ? What kind of data do we need to determine this? or is there no choice but to do load testing using production data?
Would appreciate any help. Thanks
EDIT: Additional information
Total no of records now in table is 2.3 million. Total number of records added by all the batch jobs: ave of 5000 per day