Commercially available fast-response analysers for methane (CH4) and nitrous oxide (N2O) have recently become more sensitive, more robust and easier to operate. This has made their application for long-term flux measurements with the eddy-covariance method more feasible. Unlike for carbon dioxide (CO2) and water vapour (H2O), there have so far been no guidelines on how to optimise and standardise the measurements. This paper reviews the state-of-the-art of the various steps of the measurements and discusses aspects such as instrument selection, setup and maintenance, data processing as well as the additional measurements needed to aid interpretation and gap-filling. It presents the methodological protocol for eddy covariance measurements of CH4 and N2O fluxes as agreed for the ecosystem station network of the pan-European Research Infrastructure Integrated Carbon Observation System and provides a first international standard that is suggested to be adopted more widely. Fluxes can be episodic and the processes controlling the fluxes are complex, preventing simple mechanistic gap-filling strategies. Fluxes are often near or below the detection limit, requiring additional care during data processing. The protocol sets out the best practice for these conditions to avoid biasing the results and long-term budgets. It summarises the current approach to gap-filling.