It's a TV episode. the heroes reach a low point where they think they're defeated, one of the characters has an idea to make things work, they follow him and it eventually leads to a combat scene, they lose and are hopeless again, but another character has another idea, it doesn't involve combat, just a montage of the events, and it works. Is it considered bad structure to have 2 low points? I really think both the first and the second have essential scenes that i would be sad to throw away.
I'd say yes. That first idea is a "Hope Spot", where the audience is led to believe things might work out after all, only for them not to. It's a fairly common technique (obligatory TV Tropes warning).
However, I feel like making the resolution a montage sequence is a bad idea. You're making us watch the heroes lose twice, and then not showing us how they win! So I'd do it the opposite way round: don't show the first idea's failure. Show the character saying, "I've got an idea, we can still make this work!", then smash cut to the end of the fight, where they've just been defeated again. You can still have your scene of them being hopeless and despairing, and then you can use the time you saved to show the resolution in more detail.
It is not bad structure, it is in fact common advice: Try & Fail, Try & Fail, Try &Succeed.
It is common advice to writers because many stories have succeeded using this formula. Look at the Die Hard movies or 007 movies; the villain escapes several times before they are finally caught or defeated. The same was true for 24 Hours. It is not unusual in Romantic Comedy, or other action adventure.
That said, while it is fine that another character has a new idea that does not involve battle, I strongly discourage glossing over this idea with the equivalent of "telling" instead of "showing". If anything should be visualized it is the details of bringing the reader to the final victory.
Montages are reserved for compressing a lot of boring work by the characters into something informative and clever. Preparatory work like building a new device, painting and outfitting a car to be a fake cop car, "thinking", practicing a move, rehearsing a scene, training up (think the Rocky montage), etc. You can see this kind of montage in something like Ocean's Eleven or many other heist films.
Pacing and music can be an important part; I have seen montages with every scene less than half a second long, others where the scenes are several seconds long. Scene length depends on what each scene is supposed to say; with very fast scenes the information is primarily visual; like the sights seen on a long trip. In something like Ocean's Eleven, the prep montages are slower because they are solving problems or rehearsing stuff.
The purpose of such a montage is to convey a lot of information fast in a cinematic way; but the information conveyed should still be important to the story; either character building (seeing a character moving through the world), or they are there for plausibility: Ocean's Eleven and Rocky feel more plausible because of their montages, they are "proof" of intense preparation and obstacles that have been overcome.
If that is the kind of montage you mean, then fine. But you need an extended several-minutes scene at the end, when it all comes together, for the finale.
This is because montages ONLY convey information, they are very poor at conveying emotions, and the finale must be an emotional victory of some sort, a permanent change of fortunes or something like it. The villain or obstacle is finally defeated. A 60 second montage of a man trudging across the Sahara and slowly dying is fine, but it can't include him getting out of the Sahara, that is a SCENE.
Hopefully that makes sense to you!
I think it's fine to have an up-down-up-down structure, but that one of the lows, probably the second one, should be significantly "lower" than the other one. Otherwise it does just feel like a repetition. Maybe experiencing the first low helps in some way in overcoming the second. Or the first foreshadows the second. But they should be discernibly different, and each should play a role.