Mass extinctions are selective: not all living organisms will be affected by it to the same extent. Meaning also that various groups will recover from it at various speed: a group from which half of the species were exterminated by the event will most probably take longer to recover (i. e. reach its pre-event diversity) than a group for which only 10% of the species are gone. And a group for which 95% of the diversity is gone is likely to never fully recover its initial diversity.
If you are looking at the recovery of the ecosystem/biome as a whole, then what is likely to happen is that the extinct species left whole ecological niches vacant, meaning that this vacuum is likely to be quickly filled by new species (either species group that already existed and were already occupying this ecospace but were dominated by the now-extinct species or brand new species group). On that subject, one has to realise also that the recovery of a trophic level of an ecosystem is dependent on the recovery of the others (i. e. and to simplify, recovery of apex predators depends of the recovery of grazers, which depends on the recovery of primary producers etc.), so a mass extinction affecting primarily the primary producers might need longer to recover than one affecting mainly the apex predators.
So is 10 Myr a short amount of time to recover from a mass extinction? I'll say that one would expect ecological recovery to be faster than that (see for instance the speed at which milieux disturbed by glacial episodes recover during interglacials, or the speed at which new islands get colonised). But for the complete recovery of global diversity (in particular for an event supposedly wiping out 90% of life diversity) i'd have expected it to take longer than 10 Myr. In the second article you link, the authors estimate that the recovery of the terrestrial ecosystem is fast by comparison with how long it took to colonise the land in the first place (i. e. 100 Myrs). But all the evolutionary adaptations needed to function on land do not need to be re-evolved, so it seems rational to expect it to take a shorter time than those 100 Myrs... Otherwise the probable reason the news article you linked considered it to be a long recovery time was to emphasise the fact that the article show that it took longer because of ripple effects of the Permo-Triassic event during the early Triassic that slowed down the recovery.
In a review of mass extinction recovery predating those two articles, Erwin (1998) estimated the duration of the recovery from the End Ordovician, Frasian-Fammenian, end-Permian, end-Triassic and K/T extinction events (I mention here only the so-called "Big Five" extinctions) to, respectively, 7, 3, "3 to 9?", 3 and 2 millions years. With those number in mind, then indeed 10 Myrs is a long recovery.
The same article also had this to say about the definition of mass extinction recovery:
How does one define a recovery? Mass extinctions create new ecological opportunities, and the common-sense definition of a recovery interval encompasses the rapid diversification by surviving lineages following the end of the extinction and ending with either the waning of the rapid, postextinction diversification or the return of normally functioning communities. Normal is hard to quantify, however, even without the extensive disruptions associated with some mass extinctions. Moreover, different clades recover at different rates and the same clade can recover at different rates in different regions. If we accept the recovery as beginning with an increase in evolutionary rates, this usefully distinguishes an immediate, postextinction survival interval, dominated by surviving lineages, from the often exuberant diversifications of the subsequent recovery phase. An alternative definition relies on the 're-establishment of pre-extinction isotopic values and ecosystems that are essentially equivalent to those that existed before the extinction'. Here the emphasis lies equally on the physical environment and the ecosystem recovery.
Douglas H. Erwing, 1998. The end and the beginning:
recoveries from mass extinctions, Trends in Evolution and Ecology.
Finally, evolution is non-repeatable: so to answer your question "Let's say some catastrophe strikes and wipes out everything but bacteria - are we likely to see complex life ever again", the answer is that if all eukaryotes are wiped out then no eukaryotes is likely to reappear, but another "complex" group can appear anew (though bacteria are already very complex if you ask me). Similarly if life is completely wiped out I don't think it is reasonable to think it will reappear (in particular because the Hadean/Archean conditions that saw its appearance do not exist anymore).
