Regarding pressure vessel tests, please refer to the applicable standard to your region. Typically, this is a hydrostatic test. As the name refers, its static, so, supposed to be no leakage at all. I assume that you have a vessel to test and is not able to assure its leak tightness. Am I right? Could we have some more details to check and see if more ideas arise?

That which is deemed acceptable by the engineer/surveyor of the insurance company responsible for providing burst indemnity insurance. Open a dialogue with them without delay.

Oh, the vessel has been hydrostatically tested to a higher pressure previously, hasn't it?

make sure you are measuring your pnuematic test pressure with a calibrated pressure gauge, with proper shop certification. most standard gauges can be +/- 10% or more accuracy. a test gauge is typically in 1psi (or 0.1 bar) increments and is accurate to +/- 0.5%. once test is complete, it can be removed. also, be sure to have a proper pressure procedure which brings up pressure gradually, usually in 10% increments for safety reasons. so for a max pressure of 1000 psi, bring pressure up in 100 psi increments, observe, and leak check. Once you get to max pressure, system should hold for some period of time, like 24 hours. Make sure you monitor ambient temperature, as this will affect your pressure reading. As mentioned before, these terms should be worked out with the approving entity (insurance company, state inspector)