A fluctuating process value does not necessarily mean the controller is at fault. A useful review starts with measurement and follows the actual response of the equipment.
Understand the measurement chain
Identify the sensor supplying the measured value, its installation position and how the controller processes its signal. Range, units, scaling and timestamps need to agree. An implausible reading may originate in the measurement chain rather than the process.
Observe equipment response
Compare setpoint, measurement and output over time. If a drive or valve reaches a limit, changing controller settings alone may not solve the problem. Changes belong in an approved test window with the original configuration documented.
Record the findings
Record operating conditions, the tested change and the result. This helps later shifts distinguish sensor faults, limited actuator response and control deviations. A traceable diagnosis is more useful than changing several parameters at once.