Being on an eccentric orbit, I imagine the launch windows are instantaneous (which they are) and different for each launch site whether it be Baikonur, Cape, Tanegashima, or the ESA launch site. How far apart are each launch windows?
Eccentricity has very little to do with launch windows in general, and the International Space Station's specifically. The ISS's orbit is very close to circular, with an eccentricity of 0.003 or less.
There are many factors that determine whether or not a launch window is instantaneous. A non-instantaneous launch window requires that the launch vehicle be reprogrammable to accommodate the needed trajectory changes. A non-instantaneous launch window also requires that the launch vehicle be loaded with more propellant than that needed for an instantaneous launch window and requires allocation of more time on monitoring equipment, communications equipment, personnel, etc.
If a launch vehicle successfully launches on its very first attempt, an instantaneous launch window costs the provider less in terms of vehicle complexity and in terms of other costs associated with launch. An instantaneous launch window adds risk of not making that first attempt compared with a vehicle with a non-instantaneous launch window. So it's a tradeoff.
Two key factors determine when an instantaneous launch window occurs. One is the time at which the ISS's orbital plane crosses over the launch site, heading in the right direction. This occurs about once per day as the Earth rotates under the ISS's orbital plane. The other is when the ISS passes over the launch site. The launch vehicle needs to be launched slightly before or after this occurs, depending on phase loss during launch and phase gain while in transit to the ISS.
A vehicle with a non-instantaneous launch window needs to be able to make a plane change to put the vehicle on the same orbital plane as the ISS and needs to be able to accommodate the phase change due to launching a bit early / a bit late. The required plane change is typically performed very early in the launch sequence as even the slightest delay drastically increases propulsion expenditures. The vehicle launches toward what would be not quite the correct heading so as to meet the desired orbital plane and then turns to be at the right heading. The phase change is best handled by slightly changing phasing orbits. This is a bit difficult with trajectories that have no phasing orbits.
This chart shows the launch windows for shuttle mission STS-127 (a mission to the ISS) for a one-week period.
You can see the windows were 10 minutes long, not instantaneous. There was at least a flight day (FD) 4 rendezvous opportunity every day and most days had a FD 3 opportunity. The chart is a bit confusing because the Y axis is OMS 2 Phase Angle but you can read the GMT times of the windows off the chart itself. The window open times are approximately 24 hours apart but creeping up ~22 minutes each day for the FD 3 opportunities.
Chart is from the STS-127 Level A Groundrules and Constraints
Note: there is a good explanation of the Y axis in this answer